Ideology
A Multidisciplinary Approach
Teun A. van Dijk
1
S AGE Publications
London • Thousand Oaks • New Delhi
Teun A. van Dijk 1998
First published 1998
Ah rights reserved. No part of this publication may be
reproduced, stored in a retrieval system, transmitted or
utilized in any form or by any means, electronic, mechanical,
photocopying, recording or otherwise, without permission in
writing from the Publishers.
SAGE Publications Ltd
6 Bonhill Street
London EC2A 4PU
SAGE Publications Inc
2455 Teller Road
Thousand Oaks, California 91320
SAGE Publications India Pvt Ltd
32, M-Block Market
Greater Kailash — I
New Delhi 110 048
British Library Cataloguing in Publkation data
A catalogue record for this book is available from the British
Library.
ISBN 0 7619 5654 9
ISBN 0 7619 5655 7 (pbk)
Library of Congress catalog record available
Typeset by Photoprint, Torquay
Printed in Great Britain by The Cromwell Press, Trowbridge,
Wiltshire
Contents
vii
Preface
1 Introduction
Part I: Cognition
2 Ideas and beliefs
3 Social beliefs
4 Structures and strategies
5 Structures of ideologies
6 Values
7 Mental models
8 Consistency
9 Consciousness
10 Common sense
11 Knowledge and truth
12 Identity
13 Social cognition
15
28
53
65
74
78
90
96
102
108
118
126
Part II: Society
14 Ideology and society
15 Groups
16 Group relations
17 Elites
18 Dominant ideologies?
19 Institutions
Part
135
140
161
172
179
186
Discourse
20 The relevance of discourse
21 Discourse structures
22 Context
23 Reproduction
24 From cognition to discourse
25 Persuasion
26 Legitimation
27 Ideological discourse structures
191
200
211
228
235
243
255
263
vi
Contents
28 The ideology and discourse of modern racism
277
29 Conclusions
313
Notes
References
Index
321
341
366
Preface
Most scholars have a number of dream projects in mind: topics they have
always wanted to investigate and write about, but for various reasons never
did. I have many such unrealized acadernic dreams. For many years, an
innovative study on the relations between ideology and discourse has been
one of them.
Ideology has been dealt with in literally thousands of books and articles,
but (as many other authors also conclude) its defmition is as elusive and
confused as ever. So, to write a book that would specifically deal with the
complex relations between ideology and discourse is more than a challenge:
it is pure hubris, especially since such a book should of course begin with a
proper theory of ideology. How could I possibly contribute anything new
and interesting given such an enormous literature?
Not surprisingly, therefore, it soon turned out that the theoretical component of such a study would itself be a major undertaking. One single book
would barely be enough to explore the many issues, concepts and disciplines
involved in the analysis of ideology, let alone the relations between
discourse and ideology.
Nevertheless, I accepted the challenge, and this book is the first instalment
of this major enterprise. It discusses some of the fundamental concepts of
such a new, multidisciplinary theory of ideology, and sketches the overall
outline of the ways ideology is expressed and reproduced by discourse. The
overall theoretical framework for my approach to ideology may be summarized by the triangle formed by the concepts Cognition, Society and
Discourse. That is, first, the status, internal organization and mental functions of ideologies need to be studied in tercos of social cognition. Second,
the conditions and functions of ideologies are obviously not only cognitive
but also social, political, cultural and historical. And third, ideologies are
formed, changed and reproduced largely through socially situated discourse
and communication.
Instead of simply adding results from psychology, the social sciences and
discourse studies, however, these three central concepts have to be reformulated and integrated into one theoretical framework. Discourse should be
explicitly related to the structures and strategies of the personal and social
mind, as well as to those of social situations, social interactions and societal
structures. In the sarne way, also cognition should be linked with both
discourse and society, thus serving as the necessary interface by which
social structure can be explicitly related to discourse structure.
viii
Preface
The vast majority of studies of ideology (whether Marxist or non-Marxist)
are rooted in the social sciences and pay extensive attention to ideologies in
relation to class, dominant groups, social movements, power, the political
economy or, more recently, to gender and culture. They have paid less
attention to the cognitive and the discursive dimensions of ideologies,
however. Indeed, classical work hardly analyses the details of the 'ideas',
'beliefs' or 'consciousness' assumed to constitute an ideology. Even most
contemporary approaches ignore the advances in current cognitive science,
and, vice versa, most cognitive science is barely interested in questions of
the mental structures and functions of ideologies. This is why I pay more
attention to this cognitive dimension of the theory, while emphasizing that
ideologies may well be located in the mirad but that this does not mean that
they are therefore less social.
Though usually of later vintage, extant work on discourse and ideology
does of course emphasize the important role of text and talk in the
(re)production of ideologies. To my knowledge, however, among the many
studies of ideology, some of which also deal with language or discourse,
there is not a single one that details how exactly ideology shapes text and
talk, and conversely, how it is formed, acquired or changed by discourse and
communication.
As part of the more social and political component of the theory, and
establishing an explicit link with my previous big project, this book will
occasionally use racism and racist ideologies as an illustration of the
theoretical argument. This does not mean, however, that I offer a fully
fledged account of racist ideologies, which would need to be accounted for
in a separate monograph. However, whereas throughout the book my
comments on ideology, racism and discourse will be rather general, Chapter
28 offers a concrete case study of a recent text about race relations in the
USA, namely, Dinesh D'Souzá s book The End of Racism.
A muldisciplinary theory of ideology can be accomplished only by
reducing its complexity. I am not a psychologist, a sociologist or a political
scientist. This means that my overall perspective and organizing conceptualizations will often be those of discourse and discourse analysis. There
are worse biases, given the fundamental role of discourse in the formation
and expression of ideology as social cognition and in the reproduction of
ideologies in society. At the same time, it is obvious that this book cannot
do, redo or undo the relevant substantial work that has already been done in
the social sciences.
As suggested, this book is the first result of a bigger project. It sketches
the overall framework of the theory. In later studies I hope to detall each of
its main components, namely, those of social cognition, social interaction
and societal structures, as well as the structures of discourse involved in the
expression and reproduction of ideology. These studies will also feature
concrete empirical studies of the relations between discourse, cognition and
society, as well as more detailed reviews of the relevant literature.
Preface
ix
I may disappoint sorne of the readers whose notion of discourse is
exclusively associated with the more literary, philosophical or postmodern
notions of 'discourse' or 'text' After more than thirty years, the study of
discourse has become a multidisciplinary field, featuring sometimes highly
explicit and detailed theories of structures and functions of text and talk.
Unfortunately, many fashionable approaches that currently refer to 'text' or
'discourse' ignore these advances, and for that reason offer an unsuitable
basis for a theory of ideology.
Even in a theoretical book like this, I highly value accessibility for
scholars and students from different disciplines. This mean that esoteric
jargon will be avoided, and theoretical terms only introduced and explained
where necessary. Many of the notions dealt with in this book have been
discussed in sóinetimes rather technical earlier studies. In order to be able to
construct an integrated theoretical framework, many details had to be
ignored in favour of the overall outline of the theory. I hope to be able to pay
attention to these details in the following studies in this series.
There is another way in which this book differs from much other work on
ideology: it barely looks back. Many studies, as is customary in philosophy
and sociology (and much less in, e.g., psychology and linguistics), are
commentaries (on commentaries) on the classics, from the French philosophes and Marx/Engels to Lukács, Gramsci, Althusser, Foucault and
Habermas. (For detail, see the excellent introductions and historical overviews by Larrain, Eagleton and Billig, among others.)
In this book I want to go beyond such a history and philosophy of
ideology, and integrate new ideas of contemporary discourse studies, linguistics, cognitive science, political science and new developments in the
other social sciences. In other words, in order not to get entangled in endless
debates with the classics, I have left such debates to the many other authors
who engage in them. Instead, I present a systematic, analytical study, in
which the old debates and current other studies of ideology play a role only
in the background, in the footnotes and in the references. This of course does
not mean that I think most earlier work on ideology is irrelevant. On the
contrary, there are many studie s whose theories, concepts and empirical
results are also suitable for my own project. However, in the first, theoretical
book of this project, I prefer to focus on the oyeran framework and present
that as clearly and analytically as possible, without extensive comments on,
discussions with, or references to the vast amount of earlier work. Moreover,
in order to keep the already extensive bibliography within normal proportions, most references will be to books and not to articles. In the next
volumes I hope to enter more explicitly into a debate with other approaches
to ideology.
Since this book will, I hope, be followed by others in this project on
discourse and ideology, I welcome the comments of readers. They may help
me improve the theory in these next studies.
x
Preface
Acknowledgements
Finally, I am glad to be able to acknowledge the comments on an earlier
version of this book by Michael Bil ig, Terry Eagleton, Philomena Essed
and Ruth Wodak I am especially grateful for the extensive comments by
Martha Augoustinos and Luisa Martín Rojo. Some of them kindly disagreed
with the overall perspective, others with details of my discussion. I have
tried to argue as clearly as possible where some of these discrepancies are
inherent to the choices I have made in this book, and in other cases gladly
corrected my errors and filled various gaps. Otherwise the usual formula
applies: any remaining errors are of course mine.
Teun A. van Dijk
Introduction
The fuzzy life of 'ideology'
It's almost a routine. Studies of ideology often begin with a remark about the
vagueness of the _notion and the resulting theoretical confusion of its
analysis, as I did in the Preface. Indeed, of all essentially contested and
controversial concepts in the social sciences and the humanities, that of
'ideology' may well come out near the top of the list. One historical and
political — and, yes, ideological — reason for this special status may be that
'ideology' is one of these notions that have divided Marxists and nonMarxists, as well as 'critical' scholars and 'uncriticar ones — obviously
divisions that are themselves ideological.
Still, as a general concept, ideology is hardly more vague than similar Big
Terms in the social sciences and the humanities. In many respects, the same
holds for such notions as 'society', 'group', 'action', 'power', 'discourse',
'mind' and 'knowledge', among many others. These notions defy precise
definition and seem to happily live the fuzzy life inherent in such catch-all
terms that denote complex sets of phenomena and that are the preferred toys
of philosophers and scholars in the humanities and the social sciences.
Where 'ideology' differs from these other general notions, however, is that
its conunonsense usage is generally pejorative.
Definitions generally are hardly adequate to capture all the complexities
of such notions. Indeed, such fundamental notions are the objects of inquiry
for theories and whole disciplines. Definitions cannot be expected to
summarize all the insights accumulated in such bodies of knowledge — even
if there were no controversies over the meaning of the central concepts of
such disciplines. In sum, as with many similar notions, and apart from its
uses in everyday discourse, the various versions of the concept of ideology
are simply the scholarly constructs of competing theories. That is, at least
with this word, it is as Alice was told in Wonderland: we define what the
word means. Of course, presuming that 'we' have the power to do so.
Traditional approaches
Despite the controversies and the many different approaches to the concept
of ideology, the historical terms of the debate remain remarkably similar.
We are routinely brought back to the eighteenth century, when Destutt de
Tracy in France proposed a 'science of ideas' to be called idéologie, a
Introduction
science which, incidentally, never made it, unless we take philosophy (or
psychology?) as its current representative. 1 With equal predictability, we
will then meet Marx, of course, usually in the company of Engels, and then
their followers (neo- or not) in our century, among whom Lukács, Gramsci
and Althusser play a prominent role. Similarly, on the non-Marxist side, we
are bound to encounter a sequence of sociologists and philosophers, of
whom Durkheim and Mannheim are merely the most famous. 2
As is customary in sociology and philosophy, these and other classics still
appear so prominently in most current discussions of ideology that it is hard
to find more analytical and sophisticated studies that integrate new concepts
and insights of contemporary approaches in the humanities and the social
sciences.
The remnants of the classical debates are also crystallized in the everyday,
commonsense uses of the notion of 'ideology', namely, taken as a system of
wrong, false, distorted or otherwise misguided beliefs, typically associated
with our social or political opponents. Por many in the West — laypersons,
politicians and scholars alike — communism was (based on) such an
ideology. It was often seen as the prototype of an ideology. 3 The legacy of
Marx and Engels, to whom this negative, critical concept of ideology is
usually attributed, is thus posthumously discredited by the very notion they
introduced themselves.
At the same time, this negative meaning and uses of the everyday concept
of ideology shows what most earlier analysts also emphasized, namely, that
ideologies express or conceal oné s social or political position, perspective
or interests: few of 'us' (in the West or elsewhere) describe our own belief
systems or convictions as 'ideologies'. On the contrary, Ours is the Truth,
Theirs is the Ideology. Capitalism, the Market, or Christianity, even when
'we' are no fans of them, are 'ours' and therefore not usually described as
ideologies in everyday discourse.
We see that as residues of scholarly debates, commonsense conceptions of
the notion of 'ideology' capture in a nutshell many of the main tenets of the
classical tradition: (a) ideologies are false beliefs; (b) ideologies conceal real
social relations and serve to deceive others; (c) ideologies are beliefs others
have; and (d) ideologies presuppose the socially or politically self-serving
nature of the definition of truth and falsity.
The critical element of the notion of ideology in this tradition is usually
associated with various notions of power and dominance. Following Marx!
Engels, ideologies were first of all defined as the prevailing ideas of an age.
According to the political economy of these philosophers, these dominant
ideas were associated with those of the ruling class. They are part of the
'superstructuré and hence determined by the economic or 'material' base of
society. Because the ruling class, however defined, controls the means of
production, including the means of the (re)production of ideas — most
notably those of politics, the media, literature and education — they are also
able to make their ideologies more or less accepted by the ruled as the
undisputed knowledge of the 'natural' ways things are.
Introduction
Later debates in this Marxist tradition, however, questioned the economic
determinism of the dassical definitions of ideology. Ideas, laws, philosophy,
literature, and hence also ideologies, may in part develop autonomously with
respect to the material base, and may even exercise their influence, topdown, on that economic infrastructure. 5 With Gramsci, these relations
between ideology and society were conceptualized in terms of legemony'.
Thus, instead of the imposition of dominant ideologies by a ruling class,
hegemony more subtly works through the management of the mind of the
citizens, for example by persuasively constructing a consensus about the
social order.
It is especially this neo-Marxist view and its variants that have inspired many debates on ideologies at least until the demise of Communism
around 1990, wheñ the terms of the debate changed again. Many of these
approaches are now merging with a broader critical concept of ideology, for
instance in the field of cultural studies. 6 Yet, whether as dominant or
hegemonic ideologies, contemporary versions of the Marxist idea of the
combined socio-economic and symbolic power of elite groups remain alive
in many current approaches to ideology. In my own account of the role of
the elites in the reproduction of racism, we shall encounter a special version
of that idea.
In ongoing dialogue with (and often in opposition to) the Marxist strand
of the tradition, sociologists and philosophers have continued to debate, with
increasing sophistication, the social and political dimensions of knowledge,
truth and scholarship itself. Por a long time, their insights into society were
precisely self-defined as non-ideological, and hence as truthful and scientific. Both politically and in a scholarly context, Marxism was no exception.
Aboye the fray of politics, and unbound by social or economic interests,
thus, most scholars considered themselves afreischwebende Intelligenz, that
is, beyond the pale of self-serving falsehood, and only interested in the
disinterested search for the truth — only to be accused by more critical others
of engaging in precisely what they wanted to avoid in the first place, namely,
an ideology. This ideology of science, which tries to conceal its interests and
wants its own beliefs to be accepted as truth by those who recognize its
power and dominance, is thus hardly different from other ideologies that are
developed to achieve hegemony, to legitimate power or to conceal inequality
— if only in the domain of knowledge. It is at this crucial point where the
philosophy and sociology of ideology and the philosophy and sociology of
science overlap.
It is only in a later stage, in the second part of the twentieth century, that
more inclusive and less pejorative notions of ideologies develop. Here,
ideologies are usually defined as political or social systems of ideas, values
or prescriptions of groups or other collectivities, and have the function of
organizing or legitimating the actions of the group. ' Most later work on
political belief systems is rooted in this more general concept of ideology.
It is at this point where my own exploration will start. Yet, it will be
emphasized that also the notion of a 'belief system' is still much too
1
4
Introduction
general, and in need of further analysis. This is one of the reasons why this
study also intends to continue (the few) psychological approaches to
ideology. 9
In this informal summary of some main strands of the classic debate about
ideology, most notions, including the more controversia) ones, are as
familiar as the narres associated with them. Although many are not very
precise, as may be expected for such fundamental terms, these are the
notions that are being used, and that have influenced the foundations of
virtually all social sciences. Most studies of ideology, instead of going
beyond the classics, keep repeating, reformulating and reinterpreting this
Master Narrative of the Story of Ideology. Therefore, I feel delightfully free
to presuppose Chis history to be known and to explore new ways of
addressing the old problems, and at the same time perhaps create some
interesting new problems.
A framework for a multidisciplinary theory of ideology
The philosophy and sociology of science tell us that old theories and
approaches are seldom discredited because they are explicitly proven false
or inadequate. Rather, other ideas become accepted that seem to be more
attractive for whatever reason, sometimes because they provide a better
account of the'facts', or because they focus on other, more interesting facts.
Therefore, for the same strategic reason that I presuppose the history of the
study of ideology to be generally known, it will not be my aim to discredit,
attack or debate the multitude of such classical approaches. Such a dispute
would precisely look back and remain entangled in the same frameworks of
discussion and thought (see, however, some of the endnotes for comments
on the relevant literature). Of course, Chis book cannot start from scratch, and
will use and integrate those classical ideas about ideology that remain
relevant in a new approach.
My main purpose, then, is to look ahead, to find alternative theoretical
frameworks, to explore and incorporate other disciplines, and especially to
work towards a comprehensive theory of ideology. Among other things,
such a theory would describe and explain the following.
• the general status of ideology as a cognitive and social system
• the differences between ideologies and other (systems of) 'ideas'
• the components and internal organization of ideologies
• the relations between ideologies and other shared social representations
• the relations between ideologies and values
• the relations between ideologies and social structures
• the relations between ideologies and groups and their interests
• the institutional embédding of ideologies
• the relations between ideology and power and dominance
• how ideologies are acquired, used and changed
• how ideologies are reproduced
Introduction
• how ideologies are expressed in social practices in general
• how ideologies are expressed and reproduced by discourse.
Obviously, this is a research agenda that could keep several hundreds
of scholars busy well into the next millennium, so my aims have to be
more modest, and I shall therefore focus on only some aspects of such a
theory.
I need not do so merely within the confines of the disciplines that have up
to now dominated the debate, namely, philosophy, sociology and (in part)
political science. Since I continue to talk about ideology, some of the more
familiar notions, also from these disciplines, will appear again in my own
approach. However, where necessary, they will be framed and formulated in
a novel way, and related to concepts and theoretical developments that
hitherto have received scant attention from the leading ideologues of the
study of ideology. That is, a theory of ideology first of all needs to be
multidisciplinary
Yet, we all have our limitations, interests and preferred ways of thinking,
and my approach wiil therefore be located in the conceptual and disciplinary
triangle that relates cognition, society and discourse. There are worse sites of
inquiry when dealing with the notion of ideology. First, even among those
who deny it, ideologies are at least implicitly taken as some kind of'system
of ideas', and hence belong to the symbolic field of thought and belief, that
is, to what psychologists call 'cognition'. Second, ideologies are undoubtedly social, and often (though not always) associated with group interests,
conflicts or struggle. They may be used to legitimate or oppose power and
dominance, or symbolize social problems and contradictions. They may
involve social collectivities such as classes and other groups, as well as
institutions, organization and other parts of social structure.' Hence the
pervasive interest of sociologists and political scientists in the notion of
ideology. And third, many contemporary approaches to ideology associate
(or even identify) the concept with language use or discourse, if only to
account for the way ideologies are typically expressed and reproduced in
society." Concealment, legitimation, manipulation and related notions that
are seen as the prime functions of ideologies in society are mostly discursive
(or more broadly semiotic) social practices. Of course, as we shall see, this
does not mean that ideologies are expressed only by discourse, but merely
that discourse has a specific role, among other social practices, in the
reproduction of ideologies.
Having staked out this very broad and multidisciplinary field of inquiry, it
is my contention that precisely the complex relationships involved here —
namely, those between cognition, society and discourse — are needed in an
explicit theory of ideology. To say that ideologies are systems of 'ideas' and
hence in need of a psychological approach will be an interesting suggestion
only if we realize at the same time that diese 'ideas' are also social (and
political and cultural), and that we therefore need to account for them in
Introduction
terms of the study of social representations and their functions for
social
cognition. 12
And conversely, if ideologies are part of social structure and somehow
exhibit or even control the relationships of power and dominance between
groups (classes, social formations, organizations, etc.), such a sociological
approach will similarly be relevant only if we realize that ideologies
characterize the 'mental' dimension of society, groups or institutions.
Combined then, these mutual relationships locate my theory first of all in a
joint psychological—sociological account of the social mind in its social
(political, cultural) context.
However, this still leaves us at a level of mental or social abstractions that
have no empirical grounds. We need to 'seé ideologies expressed or lived
by social actors, and 'at work' in concrete social situations, that is, in
everyday social practices. Many of these practices would do as a domain of
empirical research. Thus, forms of everyday discrimination against women
and minorities may be studied as manifestations of sexist and racist
ideologies. However, although we may well assume that such discrimination
is largely ideologically based, it does not as such 'articulate' these ideologies
themselves — at least not as explicitly as the discourses that explain, defend,
legitimate, motivate or otherwise lormulate' fragments of the'underlying'
ideologies.
In other words, although discourses are not the only ideologically based
social practices, they certainly are the most crucial ones in the formulation of
ideologies in their social reproduction. Language use, text, talk and communication (together subsumed here under the overall term of 'discourse) are
needed and used by group members to learn, acquire, change, confirm,
articulate, as well as to persuasively convey ideologies to other ingroup
members, to inculcate them in novices, defend them against (or conceal
them from) outgroup members or to propagate them among those who are
(as yet) the infidels. In sum, if we want to know what ideologies actually
look like, how they work, and how they are created, changed and reproduced, we need to look closely at their discursive manifestations.
Note that such a discourse analysis itself is multiply related to a cognitive
and a social account. Discourse meanings, inferences, intentions and many
other properties and processes of the mind are intimately linked with an
adequate account of text and talk. At the same time, it has become the
standard view in discourse studies that discourses are forms of social action
and interaction, situated in social contexts of which the participants are not
merely speakers/writers and hearers/readers, but also social actors who are
members of groups and cultures. Discourse rules and norms are socially
shared. The conditions, functions and effects of discourse are social, and
discourse competence is socially acquired. In sum, discourse and its mental
dimensions (such as its meanings) are multiply embedded in social situations
and social structures. And conversely, social representations, social relations
and social structures are often constituted, constructed, validated, normalized, evaluated and legitimated in and by text and talk
Introduction
Having sketched this rich conceptual triangle of discourse—cognitionsociety, we have a unique framework to precisely articulate the relationships
that also are needed in the theoretical account of ideology. Of course, this is
a complex project, or rather a vast paradigm for research, of which one
scholar can only design the general outline and study some smaller fragments.
Aims of this study
This book aims to contribute to such a necessarily collective enterprise. In
order to be able to emphasize what has often been neglected, my contribution will focus on the structures and strategies of discourse, social cognition
and their mutual relationships, as well as on their social embedding — and
less on societal (class) structure, or on those institutional, cultural and
political dimensions of ideology that have received primary attention in
earlier work. Of course, such an emphasis does not imply that the sociopolitical study of ideology is less fundamental.
Preparing the more specific studies of ideology and its relations to
cognition, society and discourse, this book, then, primarily aims to do some
of the theoretical groundwork. It does so by examining a number of
theoretical concepts that may be needed (or rejected) in such a framework.
This also allows me to position my own approach and conceptual analysis in
relation to current and past approaches: even new theories have historical
backgrounds, and at least need to spell out which extant ideas deserve to be
further elaborated and which ones are theoretically less fruitful. Thus,
instead of reviewing again the history of such classical notions as 'ruling
ideas', 'false consciousness', 'hegemony' or I undertaK a
conceptual analysis of these and related notions in my new framework, and
will either propose to redefine them or to leave them as history.
Obviously, such theoretical and conceptual groundwork has its own
limitations. Many of the notions discussed in this study have been the object
of impressive philosophical and social scientific treatises. Some of them
(like'knowledgé or 'group') are the object of whole (sub)disciplines. I am
unable to redo or undo all this previous work. However, I discuss some of it
briefly in a new perspective and try to relate it somewhat more explicitly to
the new notion of ideology I shall develop in this and the following
studies.
Even where earlier studies are relevant for my enterprise, their main
problem seems to be the lack of theoretical explicitness. Most crucially
lacking is a theory of the internal components, structures or organization of
ideologies. Very few of the large number of studies about ideologies ever get
down to the mundane job of describing what they actually look like. In the
same way, although most studies discuss the functions of ideologies for
groups, group members, society and culture, there is not much work that
spells out the details of such social or cognitive functions and that explains
8
Introduction
ideological structures in terms of such functions. Thus, if ideologies are
being developed to legitimate' power or social inequality, what is the
precise nature of these legitimation processes and practices? And finally, if
ideologies are expressed and reproduced, if not constituted, by discourse,
similar questions may be asked — how does this happen, what discourse
structures are involved and how exactly are these related to the social
context? In sum, although much classical and current work on ideology is
interesting and relevant also for our own discussion, their analyses usually
remain at a level of abstraction that defies detailed inquiry. It is the aim of
this book to design some of the elements of a research programme that will
try to answer such fundamental questions.
The new concept of ideology
To do this, I intend to develop a new notion of ideology that serves as the
interface between social structure and social cognition. In that framework,
ideologies may be very succinctly defined as the basis of the social
representations shared by members of a group. This means that ideologies
allow people, as group members, to organize the multitude of social beliefs
about what is the case, good or bad, right or wrong, for them, and to act
accordingly.
Ideologies may also influence what is accepted as true or false, especially
when such beliefs are found to be relevant for the group. In that latter,
epistemological sense, ideologies may also forro the basis of specific
arguments for, and explanations of, specific social arrangements, or indeed
influence a specific understanding of the world in general. Note, though, that
ideologies in this framework are not simply a 'world view' of a group, but
rather the principies that forro the basis of such beliefs. Here we enter the
perennial debate about the relations between ideology and knowledge, which
we alsó need to examine in some detall.
In most (but not all) cases, ideologies are self-serving and a function of
the material and symbolic interests of the group. Among these interests,
power over other groups (or resistance against the domination by other
groups) may have a central role and hence function as a major condition and
purpose for the development of ideologies. Ideologies thus operate both at
the overall, global level of social structure, for instance as the socially shared
mental 'monitor' of social competition, conflict, struggle and inequality, and
at the local level of situated social practices in everyday life.
The core of this new concept of ideology is not an arbitrary invention that
would take us too far from earlier scholarly as well as commonsense notions
of ideology. If that were to have been the case, we should have had to invent
a new terco altogether. Several current definitions of ideology share important elements with my own. Many authors would agree that an ideology is
something like a shared framework of social beliefs that organize and
coordinate the social interpretations and practices of groups and their
members, and in particular also power and other relations between groups.
Introduction
9
Thus, to quote just one of many such definitions by an influential scholar,
Stuart Hall defines ideology as follows:
By ideology I mean the mental frameworks — the languages, the concepts,
categories, imagery of thought, and the systems of representation — which
different classes and social groups deploy in order to make sense of, figure out
and render inteligible the way society works. (Hall, 1996: 26)
We see that many elements of my own approach already appear here: a
mental framework of beliefs about society and the cognitive and social
functions of such a framework for groups. Given his other work, Stuart Hall
would probably have no objection if we were to add to his definition that
ideologies are not limited to making sense of society, but that they also serve
to regulate social practices. In the explanation of his definition, he explicitly
refers to the role of ideologies in the stabilization (and one might add, the
challenge) of particular forms of power and dominance
The aim of this book, then, is to go beyond such definitions, and actually
spell out what exactly these 'mental frameworks' are and how exactly
(members of) social groups 'make sense' of, communicate and otherwise
interact in society on the basis of such frameworks. That is, we need not
only a definition, but also a detailed theory of ideology.
The cognitive versus the social?
One possible objection to the cognitive definition of ideology in tercos of the
basis of the social representations shared by a group may be that this
approach is too 'idealise . As will become olear later in this book, such a
critique would be misguided. Ideologies are not merely defined in cognitive
terms, but also in terms of social groups, group relations, institutions, at the
macro-level and in tercos of social practices at the micro--level. It will be
emphasized that ideologies are constructed, used and changed by social
actors as group members in specific, often discursive, social practices. They
are not individual, idealistic constructs, but the social constructs shared by a
group.
However, it will also be stressed that for a theoretically useful theory of
ideology, we should analytically distinguish between these socially shared
mental representations, on the one hand, and the social practices that are
(partly) controlled by them, or by which they are constructed. Such a
distinction is as useful as that between grammars or discourse tales and
actual language use. Hence, although a theory of ideology has an important
cognitive component, such a theory would not be complete without an
equally crucial social component. This does not imply, however, that the
theory of ideology, as is the case for traditional Marxist approaches, should
be 'materialise in the sense that it is rooted (only) in the socio-economic
base of society.
In sum, and more generally for my work, I precisely advocate a productive integration of the cognitive and the social, the individual and the
10
Introduction
collective. That both discourse and ideology are social constructs and
accomplishments should by now be a truism, and it also informs the
approach in this book. Much contemporary discourse analysis is socially (or
rather 'interactively') oriented and ignores the crucial cognitive dimension
of language use and social practices. This book will therefore focus on
cognition (and discourse) rather than on the (more familiar) social dimensions of ideologies, but that does not mean that these are less important. No
adequate theory of discourse or ideology can be developed wíthout examining the role of socio-cultural knowledge and other shared beliefs that provide
the 'common ground' of all discourse and social interaction. My point is that
these 'representations' are both social and mental.
More specifically, an exclusively social or 'interactionise theory of
discourse or ideology is unable to describe in detail how exactly societal
structures (groups, power, institutions, etc.) and even social interaction and
contexts condition the actual production and understanding of discourse, and
indeed the very participation of social actors in social interaction. If
language users share knowledge, rules or 'methods' then these should also
be made explicit in cognitive terms. The Intermediaté representations and
processes involved in these complex and detailed relationships between
society and discourse are not and should not be ignored, or mystified. We are
able to explain such relationships only if we know how language users
actually go about writing or talking, reading or understanding, and indeed
interacting, that is, by thinking and by 'making sense' of what they and their
co-participants do. This does not mean that discourse (or ideology) is
reduced to individual persons, nor to their minds. But text and talk of
language users cannot be explained without at least a serious cognitive
analysis of the minds of such language users, and especially how such minds
shape and are shaped by discourse and other social practices in context.
Obviously, cognitive science does not provide the full story about the
representation and processes that are involved in language use and the
development and the uses of ideologies. Cognitive science is unfortunately
not interested very much in social representations and ideologies, nor in
social issues more generally. With some notable exceptions, most current
social discourse analysis in turn ignores cognition, for example, because it is
afraid of psychologism, cognitivism, mentalism or individualism. None of
these -isms needs to be feared as long as one knows that discourse and
ideology are social phenomena and as long as one embeds cognition in
social contexts and society. That people think, and share their beliefs, is part
of that social Efe of language and ideology, and analysing thinking and
believing, in ,detail and explicitly, is also a task of the socially minded
scholar. Theoretically, then, there is no alternative but to integrate a social
and a cognitive analysis in the study of ideology, as will be extensively
argued throughout this book. Ignoring either the social or the cognitive
dimension of ideology will imply unwarranted reduction. This book, and
also my other work on discourse, emphatically rejects such reductionism.
Introduction
11
A critical approach
Apart from being multidisciplinary, and attempting to formulate a more
explicit theory of ideology within the discourse—cognition—society triangle,
my work on ideology also aims to be critical, in the sense of articulating an
explicit position of scholarly dissent in relationships of societal dominance
and inequality. 13 Contrary to traditional critical approaches, however, this
does not mean that the definition of ideology is limited to a concept that sees
ideology only as an instrument of domination. There are good theoretical
and empirical reasons to assume that there are also ideologies of opposition
or resistance," or ideologies of competition between equally powerful
groups, or ideologies that only promote the internal cohesion of a group, or
ideologies about the survival of humankind. This implies that, as such,
ideologies in my approach are not inherently negative, nor limited to social
structures of domination.
Does this more general conception of ideology take away the critical edge
of the enterprise, as is sometimes suggested, or prevent ideological critique?
Of course it does not. No more than that the use of the general concept of
'power' precludes a critical analysis of power abuse, as well as solidarity
with the forros of counter-power we call resistance. The same is true for the
general concept of legitimation'. Again, ideologies may be critically
examined when (unjustly) legitimating power abuse or domination, but that
does not mean that afi legitimation, as such, is negative. Most forms of
applied ethics will accept the legitimation of resistance against domination.
It would be rather arbitrary to use the notion of ideology only for the belief
systems we do not accept. What about the ideological belief systems we are
indifferent about — would we have to declare them non-ideological because
we have not made up our mind about them? Obviously, as will be argued in
more detall later, this cannot be a fruitful criterion for the use of a theoretical
concept. Thus, ideologies will only be (generally) defined in terms of their
contents and structures, as well as in terms of their cognitive and social
functions.
Such a general notion is perfectly compatible with a critical analysis of
'bad' ideologies such as those of class domination, racism or sexism, that is,
of ideologies that deny, conceal, legitimate or monitor social inequality. A
general concept of ideology not only provides a more solid framework for a
critical approach, but also allows comparison among different kinds of
ideologies, the changes of ideologies from systems of resistance to systems
of domination (or vice versa), and a more coherent and complete study of
the embedding of ideologies in social cognition as well as in social structure.
In this sense, my study explicitly continues but also fries to renew the
tradition of critical theory in the social sciences and the humanities initiated
by the Frankfurt School sixty years ago. Is
My previous major project was a study of the ways racism is reproduced
by discourse. In order to establish a link with this work, and at the same time
to have a more specific example, several chapters will make some comments
12
Introduction
on racist ideologies. These examples are merely illustrative — a fully fledged
theory of racist ideologies would require a separate monograph, with its own
theoretical framework and especially a serious empirical study of the ways
racist ideologies manifest themselves, for instance in discourse.
Organization of this study
The discussion of some of the key topics of a theory of ideology will be
organized as follows. It starts with what I consider to be the core of such a
theory, namely, the account of what ideologies 'are', where we can 'fmd'
them, what they look like, what their components are and how they are
related to other phenomena of the same kind. This will bring us to the study
of ideology as the foundation of . social cognition, and of the relations
between ideologies and other mental representations, such as values, attitudes, opinions, knowledge and mental models of events. At the same time,
such an analysis allows us to spell out the cognitive functions of ideologies.
Having established such a framework, I am able to discuss more explicitly a
number of classical concepts associated with the notion of ideology, such as
(false) consciousness, truth and falsity, common sense and (in)consistency,
among others.
Next, such an account of ideology in tercos of social cognition will be
located in a social context. That is, we need to spell out first of all what it
means exactly that social cognition in general, and ideologies in particular,
are socially shared, and indeed who or which groups have them, and
especially also why. This brings us to the analysis of the fundamental social
functions of ideologies. Such functions will probably also shed further light
on the elusive problem of the intemal structures of ideologies. Similar
questions may be asked about the discursive manifestations of ideologies in
their social contexts. Which contexts, situations, participants, institutions,
groups and group relations, or other micro or macro social structures are
involved in this 'practical accomplishment' of ideologies in discourse, and
hence in the everyday enactment and reproduction of ideologies? Which
relationships of power, dominance, resistance, competition or conflict constrain or occasion such ideologies? This framework allows us finally to
discuss in somewhat more detall the many social concepts traditionally
associated with ideology, such as those of power, domination, elites,
institutions, groups and communities.
Since these social embeddings and functions are obviously the reason why
people develop and use ideologies in the first place, I might have started
with a discussion of these social notions. In many respects this would have
been theoretically more adequate. However, given the orientation of traditional research, we know much more about these social dimensions of
ideology, so that I may first focus on the less familiar study of the cognitive
core and then locate these in their social contexts and highlight their
discursive reproduction. In other words, I first want to know what ideologies
Introduction
13
'aré, that is, what they look like, so as to better be able to study their role
and function in society.
Finally, the multidisciplinary triangle requires analysis of the ways in
which socially shared ideologies manifest themselves in a specific but
crucial type of social practice, namely, discourse. That is, I need to briefly
indicate how ideologically based social representations shared by a group
influence the actual, situated text and talk of individual social actors. And
conversely, it should be spelled out how ideologies in turra are constituted,
changed, challenged and reproduced by discourse. One of the powerful
features of such a discourse analytical approach is the theoretical sophistication of contemporary accounts of the detailed structures of text and talk.
Such an analysis allows us, among other things, to focus on the relations
between discourse structures on the one hand, and the structures of ideologies on the other hand. At the same time, together with the socio-cognitive
account, this discourse approach will be needed to discuss some more or less
'discursive' notions of traditional approaches, such as persuasion, manipuladon, legitimation, concealment, and other things social actors 'do' with
ideologically based talk and text. In other words, we here deal with the many
central aspects and conditions that define the reproduction of ideologies.
The various cognitive notions discussed in Part I are discussed in theoretical
terms, without specific reference to empirical (experimental or other) evidence. Apart from defining one major part of the theory of ideology, these
cognitive notions will, however, be 'applied' in Part III, in the analysis of the
processes of discourse production and comprehension. This means that part of
the empirical evidence of cognitive concepts may be sought in the way they
explain processes of language use. That is, apart from indexing social context,
discourse structures may themselves feature indications of underlying, mental
representations. Their analysis may thus yield rich evidence for such representations and mental processing, and thus complement evidence usually
obtained in laboratory experiments. In the later, empirical discourse studies
planned in this project we hope to do just that: show how ideologies and other
social representations control discourse structures, and vice versa. Part III
provides the theoretical frarnework for this empirical study of diese relationships.
The order of the main parts of this book is merely a research strategy and
says nothing about the order, causation, primacy or hierarchy of discourse or
cognition over society, or vice versa. It does not imply, for instance, that
ideologies as forms of social representations are'first' in the mirad before
they are 'in' society, or that 'internar structures of phenomena need to be
studied before their 'external' functions. I assume that such (discursive or
ideological) structures will often develop as a function of their uses and
functions in society. Nor do I suggest that microstructures of everyday
situated interaction should be studied before (or instead of) their macrosocial
constraints, such as group relations or institutional context.
Cognition, discourse and society are related in extremely complex ways,
in which influence and dependence are usually bidirectional, multilevel, and
14
Introduction
both cognitive as well as social. In that perspective, then, there is no point in
affirming that ideologies are first or primarily or 'really' cognitive or social.
They are essentially and crucially both. That does not mean, however, that
we need to talk about everything at the same time, or that we cannot make
analytical distinctions between different dimensions, levels or orders of
phenomena, even in an integrated, multidisciplinary study. On the contrary,
understañding of these analytically established structures and functions at
various levels of description and explanation is a necessary condition for the
development of a theory of ideology.
It should be emphasized that the chapters of this study can do no more
than provide a first analysis of sorne of the key topics and the overall
framework of a new theory of ideology. Subsequent studies, for instance
about ideology and its detailed relations to the structures of cognition,
society and discourse, will then have to develop these notions with more
theoretical precision and on the basis of empirical data
Part I
COGNITION
2
Ideas and Beliefs
Ideas
Whatever else ideologies are, they have always been associated with socially
shared ideas. First, such ideas were seen as the object of a new science of
ideology, as proposed in the wake of the philosophical movement of the
French Enlightenment. Later, ideologies acquired their negative connotation
as systems of the dominant ideas of the ruling class. Or they were defined as
the false ideas of the working class's being misguided about the conditions
of its existence. As a more sophisticated version of such 'false consciousness', ideologies were later described in terms of the persuasive, hegemonic
ideas being accepted by dominated groups as part of their common sense
about the nature of society and their place in it. And finally, beyond the
confines of an analysis of class struggle, ideologies have been viewed more
generally as any system of self-serving, mythical or otherwise deceptive
ideas defined in contrast with the trae ideas of 'our' science, history, culture,
institution or party.
Whereas several of these defining notions will be dealt with later, let me
first examine what exactly these 'ideas' are. The notion of 'idea' is one
among many in the history of the study of ideology that hardly are specified
in more detail than the everyday, commonsense meanings of these terms. If
we assume for a moment that ideas (apart from being abstractions or social
constructions) are at least also things of the mind, and that therefore
psychology should tell us something about them, a relevant literature review
would be disappointing. Modem psychology books do not talk about ideas,
at least not explicitly and not in these terms: the term does not appear in the
subject índex of most current books in cognitive psychology. So let me
begin by analysing some of its everyday meanings:
1 Ideas are objects or processes in/of the mind.
2 Ideas are the products of thinking or thought.
3 Ideas are part of knowledge.
4 Ideas may be personal or socially shared.
16
Cognition
5 More specifically, ideas are new, original interesting thoughts and about
important issues.
Many standard expressions and other forms of everyday talk provide the
evidence for such conceptual meanings. People talk about ideas they have
been 'walking around' with without as yet speaking about them, as ideas that
are developing in their leads' or 'minds', as having or not 'having' an idea
(sometimes meaning that they know or doñ t know something, as in '1 have
no idea), but also as the ideas (shared by the members) of a group, a
movement, philosophers, a revolution, and so on. Thus, people may'come
up' with an idea, or an idea may'come up' with them. Conversely, we may
claim to have 'giveñ her that idea, or 'put that idea in her head'.
Often, the concept of an 'idea', whether of a person or a group, is not
merely identified with any trivial products of 'thought' one may have, but
with more original ones — the expression'I have an idea' therefore means
something like 'I have a new, original, thought.' And the hapless scholar
being accused of 'not having any ideas' is thereby damned as being someone
who has no original scholarly thought. Therefore, a'system of ideas' is
sometimes simply equated with socio-cultural, philosophical, artistic or
scientific thought or theories, as is most obviously the case in the listory of
ideas'
On the other hand, people may have 'wrong ideas', and are then accused
of ethically doubtful or socially unacceptable beliefs, and similar connotations seem to be at work when warning people not to 'get any ideas'.
These and many other colloquial uses of the concept of 'idea' clearly
signal that whereas psychology spurns the mundane notion of 'idea', the
commonsense uses focus on ideas as a specific category of (products of)
thinking, namely, fresh, original, new and sometimes unacceptable thoughts,
both in everyday life, as well as entertained by people hired to do so, such as
philosophers and other scholars, writers and artists, and indeed by 'ideologues' in the more political realm. These ideas may be expressed by the
person who has them, conveyed to others, shared by others and a whole
group; they may be further developed, influenced and manipulated. Once
shared, ideas may thus become part of the public domain, and thereby
acquire a more social or cultural dimension.
. 1
Minds
This relative vagueness of the concept of 'idea may have kept psychologists
from adopting it in their theoretical vocabulary (using instead several
notions that are barely less precise, as we shall see), but its intuitive
meanings clearly suggest that ideas are constructs or products of thinking,
that is, of the mirad, whether or not they are socially or culturally shared
Tbus, if ideologies have anything to do with ideas, then at least one of their
dimensions should be accounted for by the theories being developed in the
new cross-discipline now commonly called'cognitive science', featuring
Ideas and beliefs
17
cognitive individual and social psychology, cognitive sociology, cognitive
linguistics, philosophy, logic and artificial intelligence.
Mind versus body?
This will also be my first step: whatever else they are, ideologies are sets of
specific ideas and hence 'mental' objects. Although trivial for most cognitive
scientists, such a first step is not without controversy for some social and
discursively oriented psychologists and social scientists. For them, talking
about the 'mind' is like talking about the 'sour some centuries ago, namely,
a remnant of scholarly or religious myths, in this case of the old Cartesian
dualism separating 'mind' from 'body'. 2
This book will not waste many words on this controversy. The modem
study of cognition assumes no such duafism. As far as psychological and
neuroscientific insights go, the mind is a specific property of the brain-inthe-body. As with most psychologists, I abstract from the neurological basis
of these 'mental' properties of the brain and conduct my analysis at another
level of description and explanation. The dominant (and often contested)
metaphor of what such a mind does is that of'information processing'.
Though limited for several reasons, the metaphor has proved quite successful in accounting for at least some aspects of the typical things people are'
able to do due to their minds: perceiving, understanding, thinking, remembering, speaking and interacting. We shall later see that such brain-based
minds of persons also have a social dimension, being the product or
construct of social interaction, in their acquisition, development and uses. 3
However, this biological basis of the mind does not mean that talking
about and analysing the mind and its properties needs a reduction to the
neurobiology or, further down, to the biochemistry or the physics of neurons
or brain cells. No more so than that talking about action requires analysis of
muscle movements (and further down to the molecular and atomic properties
of nerve and muscle tissue). And no more than that a discussion of discourse
is pointless unless based on references to our articulatory or auditory organs,
air waves, the chemistry of ink or the electromagnetic properties of
computer disks.
That is, all these attempts at reduction that occasionally plague scholarly
inquiry are usually no more than a form of sometimes well-intentioned but
naive fundamentalism. They ignore both the commonsense and the scholarly
need to understand and theorize about reality at different levels or dimensions of observation, experience and thinking, including about abstractions
and things which that same mind construes for us as-if-they-were real, such
as ideas, actions, persons, groups and society itself.
In that sense the mind is a product of itsel£ And a very handy construct at
that, multiply used in everyday life, as well as in all scholarly endeavours.
Thus, when we need to talk about things like ideas, it is simply quite
convenient to do so in terms of properties of the concept of mind, whether
minds 'really' exist or not. Reification here simply is no more than an
18
Cognition
inevitable but useful product of our understanding, as long as it aliows us to
describe, explain and otherwise account for events and phenomena we want
to understand. Minds, thus understood, are both the 'means of production' as
well as the 'product' of mental activities like thought. This is what analysis
and theorizing are all about.
It is also in this cense that we accept to be 'mentalist', as long as that term
is not meant to imply that, conversely, all phenomena that have a mental
dimension are 'in fact' or 'really' only things of the mirad. Persons, actors,
actions, interactions, situations, groups and societies as a whole may be
mental constructs or have mental dimensions at some level of analysis, but
obviously a theory of such constructs needs to go beyond a'mental' analysis
and move to another level of commonsense thinking and theorizing which
we call'social'.
That I used several paragraphs to discuss the very relevance of the notion,
and hence of a theory, of mirad is merely because, as suggested, there are still
scholars who for various, sometimes somewhat (neo)behaviourist reasons
assume that minds can be dispensed with, that all allegedly mental things are
but a widespread, commonsense and psychological illusion, and that all
relevant mental notions could and should better be accounted for in terms of
what people observably do or accomplish, especially jointly, in social
situations. Since this point of view, which may be defended in more or less
radical versions, also touches upon a theory of ideology, we will have to deal
with those Interactionist' — as one may call them — ideas (sic!) later. 4 Note,
though, that my critique of anti-mentalist discursive psychology does not
imply that I disagree with much of its criticism of contemporary mainstream
psychology, such as its neglect of the socially situated and discursive
dimensions of the development and uses of 'mental' objects.
Beliefs
Although informally perfectly acceptable as a concept that may be used to
theorize about ideologies, I shall nevertheless abandon the notion of 'idea'.
Not only because it is too general or too vague, but also because it has
associations that I do not want to take along in my discussion: for example,
that ideas are often seen as new or original thoughts. Instead I shall use
another general notion of psychology, namely, that of beliefs.
Knowledge and beliefs
Many of the things that have been said aboye about ideas also apply to
beliefs. They are also products or properties of thinking, and therefore also
associated with the mirad Anything that can be thought is here taken as a
belief. However, I use the term as a technical term. This means that some
commonsense meanings of the term will not be included in the concept. For
instance, in everyday language, the concept of'belief is mostly used as
Ideas and beliefs
19
being opposed to that of 'knowledge'. Beliefs in that sense are subjective,
and may hence be wrong, unfounded or misguided. Knowledge, on the other
hand, is (the product of) thought that is found to be true.
We shall have to come back to this distinction, because it has been crucial
in the history of the study of ideology. At the moment, however, all products
of thinking will be declared to be beliefs. In other words, beliefs are the
building blocks of the mind. Knowledge in that case is merely a specific
category of beliefs, namely, those beliefs Ve' (as a group, cornrnunity,
culture, instance or institution) take to be 'true beliefs', according to certain
grounds or (truth) criteria. These criteria establish that the beliefs (for us) are
valid, correct, certified, generally held, or otherwise meet socially shared
standards of truthfulness. Obviously, these criteria are sociálly, culturally
and historically variable, and so is the knowledge based on them. This also
means that beliefs in this technical sense are not merely subjective or even
unfounded or untrue products of thought, or beliefs (like religious ones) that
are only accepted as 'true' by a specific group of people, but also include
what we call knowledge. The epistemology and psychology of these beliefs,
as constituents of ideologies, will be one of the aims of the rest of this part
of this study. 5
Judgements and opinions
Similarly, beliefs are not thoughts that are limited to what exists, or what is
(or may be) true or false. They may also pertain to evaluations, that is, to
what we think (find) to be good or bad, nice or ugly, permitted or forbidden,
acceptable or unacceptable, and so on — the products of judgements based on
values or norms. Such beliefs are commonly called opinions, to which I shall
turn later, because ideologies prominently feature such opinions. The
distinction between knowledge and opinion goes back to the classical
distinction between epistémé and doxa, made by Plato, defined as systematic
(scientific, philosophical) knowledge, and (possibly erroneous) popular
belief, respectively.
Whatever their differences, I shall provisionally subsume knowledge and
opinions under the general category of beliefs. Thus, that a specific drug has
a specific chemical formula is a belief (which we may hold to be true), and
so is the belief (which we may or may not hold to be true or defensible or
appropriate) that such a drug is good or bad for our health, or the belief that
drugs should, or should not, be prohibited, or the belief that the prohibition
of drugs, and not drugs themselves, wreaks social havoc. In sum, also the
ethics and aesthetics of the products of judgement are part of a general
theory of beliefs. Obviously, this is merely a first delimitation of the concept
of 'belief we use: specific cognitive deories provide the details of such
(still very vague) approximations. I hola such cognitive theories of beliefs to
be as necessary for a theory of ideolo&y as theories of power, group or class
in more traditional approaches to ideoiogy.
20
Cognition
However, this first approximation towards the notion of lelier does not
mean that there are no complications. First, we may as sume that people have
beliefs that are not the product of conscious thought. In the same way as
grammars of natural language are a forro of largely implicit knowledge,
people may also have many beliefs they are not aware of, and that have been
acquired without much conscious processing. This means that I do not limit
'thought' to conscious mental processes, although people may usually
become or be made aware of the beliefs they hold.
Second, we need a lower and an upper limit on beliefs. A religion might
be described as a lelief , but in such cases one should rather speak of a
belief system. This means that we need some notion of a'basic belief of
which more complex beliefs or belief systems may be composed. Ignoring a
vast philosophical discussion here, I shall simply define a basic belief as any
product of thought that cannot be decomposed into more than one belief (see
below for further discussion). Traditionally, such a belief is described by a
proposition consisting of one (n-place) predicate and n argumenta, possibly
modified by a number of modalities. Note, though, that this is a logicophilosophical definition, and not a cognitive one. The cognitive characterization may be in terms of the ways beliefs are represented, that is, in terms
of relations between nodes in a mental network, or more substantially as any
elementary thought that may be (found or made) true or false or with which
one may agree or disagree. That is, the concept of is not a belief,
whereas the thoughts expressed by the English sentences,'This is a table',
'The table is red', The flowers are on the table', and 'The flowers should be
on the table', would be beliefs. This is admittedly quite elementary, but it
will have to do for our discussion until we deal more explicitly with mental
structures.
Emotions
There is a class of 'mental' objects that may or may not be beliefs,
depending on oné s theoretical position, namely, affect or emotion. Under
one analysis, emotions are not mental at all, but constitute a separate realm.
Feeling angry or jealous, in that case, is not a belief, but at most a 'state of
mind', or even a'state of body' — for example, the tendency to hit or bate
someone. However, whatever emotions are exactly, and granted that they
are not merely of, or based on, the mirad, they do have obvious mental
(thought, belief) dimensions as well. Feeling angry or worried about
genocide in Bosnia implies or presupposes the belief that there is genocide
in Bosnia, and usually also that genocide is wrong. In this respect, emotions
may involve the (mental) interpretations of our 'state of mirad' or 'state of
body'. That is, an emotion usually has an object (though this may be very
vague), namely, what we are moved about, and if we know what that is,
emotions and beliefs need to be closely related. Hence, at some levet of
analysis, also emotions or affect belong to the realm of beliefs. This will
Ideas and beliefs
21
again be crucial for a theory of ideology, because many ideologies are often
seen to embody affect, as is the case for 'ideologies of bate' such as racism,
or 'ideologies of love' such as some religions, or the 'ideologies of anger'
that fuel resistance or revolutions. Of course, this is still highly impressionistic, so further conceptual analysis will be necessary.
Beliefs and cognition
Within this huge field of 'belief research', then, I shall first of all focus on
the cognitive and social psychology of beliefs, and later deal with their
discursive and social dimensions. In psychology the analysis of beliefs as
products of thinking, as we have seen, locates them in the mind, and more
particularly in what is called 'memory' . Memory, in this technical sense, is
nothing but a theoretical construct of the 'cognitive' part or dimension of the
mind, that is, the theoretical location where information is stored and
processed. In that sense, beliefs may be defined as units of information and
information processing, just as much as they may be seen as the products of
thinking, or indeed as the (mental) conditions and consequences of discourse
and social interaction. It all depends on the level, the scope and the nature of
the theory. Modern cognitive psychology has adopted the useful (but
limited) metaphor of information processing, without, however, implying
that our minds function like computers.
I woñ t go into the details of the properties of human information
processing, such as those involved in perception, interpretation and storage
of sense data, or the activation and uses of earlier stored units of information. I shall simply assume that beliefs are units or representations that result
from and are involved in the processes of information processing taking
place 'in' memory. In sum, the mind, or memory, is a storehouse of beliefs,
and at the same time is defined by the mechanisms (processes, strategies,
mental activities) that produce or process such beliefs. Tbus, beliefs may be
constructed, stored, reactivated, organized in larger units, and such processes
take place in the accomplishment of all cognitive tasks (which in turn are
usually part of social action and interaction).
This does not mean that all of the mirad or memory is filled with beliefs as
we want to define them. Memory may also feature information of a
structural nature (such as the composition of a sentence or a story) that is
not, in my sense, a belief. Besides such more abstract knowledge or
competence, we 'have' abilities like knowing how to walk, eat, or ride a
bike, and these are again not beliefs as intended. Beliefs apparently need
some kind of'contene or'object'. They must be about things. We believe
that something is truthful, graceful or hateful, whether or not such 'objects
of thought' correspond to something which we hold to be 'real' in the world.
We also have beliefs about 'unreal' or mental objects, such as fantasies,
dreams, goals or theories. Thus, for us, both 'thinking that' and 'thinking of
involves beliefs.
22
Cognition
Propositions
When described, beliefs usually are assigned a format of the kind 'X is (or
has the property) P', or 'X and Y are related by relation P'. Now, this is
quite similar to the kind of format we know from philosophy and logic, and
which we call propositions. It is therefore quite common to describe mental
units such as beliefs in tercos of propositions. This does not mean of course
that we actually do have propositions in memory, the mind, let alone in the
brain, but only that our theoretical language in which we speak about the
mind uses the concept of proposition in order to describe and analyse beliefs
in tercos of a propositional format. If we had a more useful format, we would
probably soon dump the notion of proposition, since it has all kinds of
drawbacks — it was especially designed to account for the analysis of what
people'pro-posé, that is, of statements and argurnents in natural language.
Although widely used in semantics for the description of (meanings of)
discourse, propositions as we know them are not exactly flexible instruments
to account for all structures of meaning. The same is true for the description
of beliefs. As we shall see later, meanings are a type of belief, namely, the
specific belief(s) associated with expressions (utterances) in natural languages. Yet, with all its limitations, I shall occasionally use this propositional format to describe the contents and the structures of beliefs. The
advantage of such usage is that propositions can again be expressed in
natural language, so that we can use natural language to talk about the
contents and the structures of beliefs.
Thus if we have the belief that the genocide in Bosnia should be stopped,
such a belief may be propositionally described as follows:
1 Must (X (stop, (commits (Y, Z(genocide)))))
of beliefs we adopt.
or similar variants, depending on the kind of
Many aspects of this belief are not part of this proposition — for example, our
knowledge about genocide, about who is responsible for the genocide (in
order for that person or group to be stopped), that the 'must' here is a moral
or political obligation, and that X is probably a powerful agent (person,
group or state), that the action of stopping X committing genocide needs to
be taken now or as soon as possible, and so on. That is, both as to content,
and to structure, a belief may be quite complex. In principie, however, all
these aspects may again be represented propositionally, so that most beliefs
are in fact a complex cluster of more elementary propositions, or simply a
propositional complex'.
This abstract language to describe mental objects such as beliefs is not
always necessary for more informal or higher-level theorizing, so I shall use
it sparingly. That is, many beliefs can simply be described more informally
with our natural language expressing such propositional complexes, such
as:
2 The genocide in Bosnia must be stopped.
Ideas and beliefs
23
3 Someone should prevent the genocide in Bosnia.
4 The massive killings of innocent people in Bosnia must be halted.
These examples also show something we shall later encounter in more
detall: that natural language expressions of'underlying' mental beliefs may
take many forms and variations. These may depend on the variable properties of the context, such as participants and their roles, goals, setting, shared
knowledge, and so on. That is, language use or discourse do not merely
express beliefs, but also are forros of action and interaction, and these
properties also influence the structures of sentences. We should merely be
aware of the fact that the sentences in 2-4 are merely expressions of beliefs
and not the beliefs themselves.
Networks
Although propositions are quite common to represent units of the mind like
beliefs, there are of course alternatives, some of which are suggested to be
closer to the neural, network structure of the brain. 8 Thus, we may represent
a belief as a collection of nodes related by paths or more specifically as
graphs with edges and nodes, and so on. The node 'genocide' may thus be
connected with the node 'Bosnia', whereas the latter node may again be
'related to the nodes 'country', 'ex-Yugoslavia' and'Muslims', whereas the
node 'genocide' may be related to such nodes as, 'mass killing','innocent
peoplé,'Holocause,'ethnic groups' or 'ethnic cleansing'. Such representations of beliefs in tercos of graphs, or further 'clown' as neural networks,
show more clearly than propositions that the'contene of a belief may be
complex, and that beliefs may be related to many other beliefs (such as,
'Bosnia is a country in former Yugoslavia', etc.). A network may then be
equivalent to a list or organized schemata of propositions, but it shows the
relevant relationships between the concepts of these propositions more
clearly. Moreover each link between nodes (where each node represents,
e.g., a neuron or cluster of neurons) may be given a certain weight or
strength of its connection depending on how often it has been activated or
used. Each belief could then be defined (at least at one level of representation) as the complex state in which the brain is when the relevant links have
been established or computed. According to such a connectionist approach,
thus, beliefs or complex belief structures are not simply located at one
specific location of the brain, but rather represented as a distributed network
of nodes and their positively or negatively weighted links In this book,
however, we shall not further explore these different modes of representation
and mental activity. Instead, we conduct our analysis at a more abstract,
macro-level of mental representation and operation, in which somewhat
easier to handle, 'symbolic' instruments such as propositions can be used.
Furtherproblems in the definition of 'belief
Thus, I shall provisionally use propositions to represent beliefs. It should be
borne in mind, however, that since we can only write or talk about
24
Cognition
propositions by expressing them in a natural language (or other sign system),
the beliefs they describe are conceptualized in terms of (the meanings of)
that natural language. This linguistic bias obscures the fact that beliefs may
be complex mental structures of which only some concepts are captured by
the propositions as expressed in sentences of a specific language. In other
words, beliefs (and the propositions we use to describe them) should not be
confused with their contextually or linguistically variable expressions.
Whether or how at least some concepts or beliefs are formed or structured as
a function of their verbal expression in a specific language, is a famous but
different problem.
The theoretical ambiguity of the notion of belief also appears in the
possibility of describing beliefs at different levels of abstraction, as we also
know from the theory of semantic macrostructures in discourse. That is, we
may at a very high level of abstraction characterize a large cluster of beliefs
about the present situation in Bosnia as a 'civil war' or as 'genocide'. Such
concepts and the propositions formed by them are, however, some kind of
high-level 'summaries' of a large number of more specific, more detailed, or
lower-level beliefs, for instance about bateles (and their details), about rape,
murder, arson, and many other acts that, together, define a civil war or
genocide.
This raises the question whether we may or should speak of basic beliefs,
that is, beliefs that do not 'surnmarize' more specific beliefs. This question is
related to the nature of thought and perception, for example, of (basic)
events, actions or properties. In the same Bosnia example, thus, we may
observe or think of people who shoot (at) others, and maybe we might
'decompose' such a perception of or thought about the act into a sequence of
components, such as laking the gun', 'raising the gun', 'pulling the trigger',
and so on. However, there seems to be a culturally conventional, basic level,
below which events and actions are no longer decomposed as 'natural' units
in everyday perception, propositions and descriptions. Thus, we do not
usually decompose and conceptualize the — theoretically infinite — movement
of 'raising the gun' in increasingly smaller parts of the movement. That is,
we may be able to actually 'see' very small movements, but they are no
longer culturally coded in a separate concept. 9 We may assume that such
conceptually driven (and probably culturally variable) perception and
thought also provides the basic level of the formation of beliefs. Most of our
beliefs about Bosnia will be at a much higher level of conceptualization than
this basic level, which is usually limited to personal experiences and direct
observation in specific contexts. For later recall, use and description, such
very low-level beliefs will usually no longer be accessible; they tend to be
subsumed by higher-level beliefs. We shall tater see that Chis also is true for
discursive descriptions, which, depending on context, genre and various
other constraints, may be fairly low-level (detailed, specific) or more or less
high-level.
That is, beliefs, whether described as propositions or networks or in terms
of other languages of (mental) representation, obviously do not come alone.
Ideas and beliefs
25
Simple beliefs may be combined to compound beliefs (such as, 'If the
leaders of large countries do not decide to do so, the genocide in Bosnia will
not be stopped'). They may cluster with large numbers of other beliefs, and
thus form the complex belief clusters we call knowledge or attitudes. Thus,
all we know about the situation in Bosnia is such a cluster of beliefs, and all
we know and think about genocide or how to prevent it, is another cluster of
beliefs.
Another issue involved in the characterization of beliefs is their relationship to the externa! world. As mental objects of some kind, they are often
taken to 'represene some'fue in the real world. Depending on our
ontology, however, such a representation—relation may have a more passive
or more active nature. If we assume facts to exist independently of the mind,
as would typically be the case for the facts (events, processes) of nature,
beliefs would rather be mental stand-iris (models, symbols, icons, images,
etc.) of the facts. On the other hand, we may also take a more active view of
beliefs, and define them in terms of socially based, mental constructs that
constitute the 'facts', typically so of social and cultural 'reality'.
I take this latter, constructive view of beliefs — representing the world,
even the facts of nature, involves interpretation and understanding of that
world in terms of socially acquired conceptual categories. In that sense
beliefs constitute the world-for-us. This obviously does not mean that the
natural or social world does not exist independently of our beliefs, but only
that people structure, understand and experience it (directly or through
instruments) in terms of their beliefs. Nor does it mean that people' s
commonsense experiences of the out-thereness of their perceived and lived
world is misguided, but only that such experiences themselves are mental
representations. What happened in Bosnia was all too real. But conceptualizing 'what happened' as a 'civil war' is obviously a mental as well as sociocultural or political construct.
In this sense, then, beliefs may still be described as being about the
objects, properties, events, actions or situations of this 'externa!' world, as
long as we realize that such an experience presupposes a socio-culturally
controlled 'projectioñ of beliefs. And for the same reason it still makes
(common as well as theoretical) sense to talk about true and falce beliefs,
depending on whether or not their representation corresponds to the 'projection tales' or truth criteria accepted within a given culture.
Apart from thus associating mental representations with the intersubjectivity of culture and society, such a constructive cognitive approach also
easily explains imaginary, fictional or abstract beliefs, lies, plans, expectations, hopes, illusions, as well as personal or social biases in the perception
and understanding of the world. Obviously, this is what we need in a theory
of ideology.
This brief account of beliefs and their propositional or other forms of
representation also suggests that even for such a fundamental notion in the
cognitive and social sciences as 'belief our theoretical frameworks are as
yet quite primitive. It is one of these notions we all use quite frequently, but
26
Cognition
when pressed to define exactly what they mean, most scholars probably will
soon give up. Because of their conceptual discreteness, propositions
expressed in some natural language at least have the advantage (and
sometimes the disadvantage) of'freezing' vast networks of conceptual nodes
into a relatively simple format. Obviously, this is also fundamental for
everyday communication, sine it is usually impossible (and mostly contextually irrelevant) to cónceptualize and express all we believe about some
situation.
Ideologies as beliefs
The reason I have talked about ideas, beliefs, compound beliefs and finally
about clusters of beliefs as constructs of the mind a is obviously that
ideologies are just that: clusters of beliefs in our minds. That is, one way to
describe and analyse ideologies is in tercos of a cognitive psychology of the
internal structures, relations, processing or other 'mental manipulation' of
(some kind of) beliefs. This is not merely a psychological trick to explain
away or reduce ideologies to memory units, but quite close to our commonsense notion of ideologies as 'systems of ideas', such as feminism,
socialism, racism, anti-racism or environmentalism. That is, we know that
feminists, socialists, and so on 'hold' or 'stand for' a number of beliefs,
beliefs about what is true or false (about gender or class relations), and what
they'fmd' good or bad (about there relations) and what should be done
about it.
Again, to account for ideologies in tercos of beliefs and belief systems,
and hence as properties of the mirad, does of course not imply that ideologies
are only mental, nor that their analysis should stop there. It has been stressed
already that ideologies are also socially shared and related to societal
structures, an obvious insight which, however, needs a different theoretical
analysis. Similarly, beliefs are not only personal, nor do they always
spontaneously'emergé as products of the individual mirad. Rather, many of
them are socially acquired, constructed and changed — for example, through
social practices and interaction in general, and through discourse and
communication in particular. This means that besides their mental dimensions, they have social dimensions, neither of which can and should be
reduced to the other. The point of any explicit theoretical analysis is to
distinguish between the different (mental, social, cultural) dimensions of
ideas and ideologies, and then to establish relationships among them. 11
Many contemporary approaches to ideology emphasize that ideologies are
not merely systems of beliefs, but also feature such phenomena as symbols,
rituals and discourse. 12 It may be readily agreed that such phenomena are
often part of ideological systems and practices in a broader sense. However,
it is theoretically more useful to distinguish between ideologies as such, that
is, socially shared beliefs of a specific type, on the one hand, and their
expression or enactment in symbols, rituals, discourse or other social and
cultural practices, on the other hand.
Ideas and beliefs
27
Of course, this again raises the broader question about the relations
between language and thought, and whether beliefs — as defined here —
presuppose language (or other forros of semiotic expression) in order to be
conceptualized. This more general question, however, is beyond the scope of
this book. With most psychologists I shall simply assume that (although of
course the mind, and hence our beliefs, is largely also acquired by language
use) specific beliefs do not themselves require a natural language in order to
be formed and used in thinking.'
In Part III, we shall fmd that discourse analysis provides 'empiricaP
evidence for the theoretical relevante of the cognitive notions introduced in
this and the following chapters. That is, although theoretical analysis of
belief systems and other mental representations may be a valid aim in itself,
a multidisciplinary theory of ideology studies such beliefs primarily in order
to describe and explain social practices in general, and discourse in
particular.
Cognitive analysis but no reductionist cognitivism
In this chapter, we have begun to analyse some of the 'mental' aspects of
ideologies, and we shall continue that analysis in the other chapters of this
first part. That is, against reductionist theses that aim to redefine ideas,
beliefs or ideologies only in tercos of social interaction or discourse, I claim
that the mind needs analysis in its own terms. But against cognitivist
reductions that claim that all social interaction and discourse, as well as
social structures, are 'really' constructs and hence products of the human
mind, I shall similarly take a social position and claim that beliefs and
ideologies also have an important social dimension that requires analysis in
its own tercos.
Thus, trying to make explicit both the commonsense notion of ideology,
as well as the traditional concept of ideology in philosophy and sociology, a
cognitive approach may spell out in somewhat more detall the componente,
contents and structures of ideologies. At the same time, it makes explicit the
relations of ideologies, as systems of specific beliefs, with other types of
beliefs, such as attitudes, knowledge and opinions. Doing so, I am sketching
the first part of a multidisciplinary framework, and designing the relevant
theoretical concepts that allow us to talk about ideologies and their embedding in cognition in a somewhat more sophisticated way than has been done
before in traditional work on ideology. This will be the task of the other
chapters in this part of this study.
9
Social Beliefs
Personal versus social beliefs
We have seen that ideologies first of all may be defined in terms of beliefs
and that such beliefs may be organized in various ways. In several
disciplines, and especially in the study of political cognition, 'belief systems' have been the standard way to talk about ideologies. 1
One of the problems with a general term such as lelief system' is that it
is too comprehensive to describe the specific sets of beliefs which I want to
call ideological. As we have seen, everything people think may be called
beliefs, and we therefore need to make further distinctions. Thus, the beliefs
expressed in the following sentences are not typical for what we usually call
ideological beliefs:
1 Water freezes at O degrees centigrade.
2 Amsterdam is the capital of the Netherlands.
3 Last month I lectured in Valparaiso.
4 I like ice-cream.
5 Krzysztof is my neighbour.
6 There was a young girl of Nic'ragua / Who smiled as she rode on a
jaguar.
That is, we knów or believe many things in everyday life that hardly fall
under the broad term of ideological beliefs, where the latter category
comprises beliefs that somehow have to do with a special ideological
'positioñ or with group interests. Among such 'non-ideological' beliefs is
knowledge about undisputed facts, as in 1 and 2, past experiences as in 3,
personal preferences as in 4, mundane facts of everyday life as in 5 and
fictional or literary'facts' as in the first two Enes of the limerick in 6. y Note
though that, as usual for isolated examples, we should add a provision like
'under standard interpretatioñ . It would not be too difficult to construct a
context or text in which even these beliefs may be ideologically based. This
is especially the case for 'undisputed facts', a commonsense category that
itself is based on a specific field of knowledge and truth criteria. Such beliefs
may be challenged (and sanctioned as 'ideological') by others, as Galileo
Galilei found out several centuries ago in bis dispute with the Catholic
Church.
Social beliefs
29
Episodic and social memory
In order to make our next theoretical step in the analysis of ideologies, we
therefore need to make a distinction between different kinds of beliefs. In
psychology such a differentiation may be associated with different regions,
parts or functions of memory, such as episodic and semantic memory.
Episodic memory is the part of memory where beliefs are stored about
concrete episodes (facts, events, situations, etc.) we have witnessed or
participated in ourselves, or about which we have information through
discourse from others. That is, episodic memory stores our personal experiences, and might therefore also be called 'personal memory'. Examples 3,
4 and 5 aboye are examples of personal beliefs as stored in episodic
memory. 3
Note that such tercos as 'episodic' or 'personal' memory are merely
theoretical constructs to account for different kinds of mental processes and
representations and their functions. As explained in the previous chapter,
such theoretical dornains of memory need not correspond for instance to
different regions of the brain (although they might — as observed for instance
by selective memory loss of personal experiences, resulting from brain
lesions). This is more generally true for the cognitive notions used in this
chapter and this book: they do not pretend to reflect the neurological or
biological properties of the brain, which would require a very different level
and kind of theorizing. 4
There are also beliefs we typically share with many others, for instance
most other members of a group, organization or whole culture, and which
therefore may simply be called social (or sociocultural) beliefs. Our vast
'knowledge about the world' is constituted by such socially and culturally
shared beliefs. These are usually located in what cognitive psychology calls
'semantic memory'. However, we shall speak of social memory, since not all
of this knowledge has to do with the general meanings of words, and hence
need not be called 'semantic' in any standard meaning of that terco.
Examples 1 and 2 aboye are typical instances of such socially shared
beliefs. 5
Ideologies, as I shall discuss in more detall later, typically belong to the
realm of social beliefs, and are therefore located in social memory. Thus, if
ideologies are belief systems, we need to be at least a bit more specific and
say that they are social belief systems.
That is, the theory being developed here emphasizes that there is no such
thing as a purely individual or personal ideology. 6 Ideologies are essentially
social, that is, shared by members of groups or collectivities of people. 7
Later (Chapter 15) we shall discuss this social basis of ideologies in more
detall, and try to find out what kind of groups typically develop ideologies.
Thus, intuitively, the people waiting at a bus stop are not the kind of 'group'
which we normally assume to share the same ideology because of the mere
fact that they are waiting for the bus together. Of course, they may
accidentally share an ideology, but not as would-be bus passengers. On the
30
Cognition
other hand, people participating in a demonstration are a more likely
collectivity to share an ideology, namely, the very ideology that in fact led
them to participate in the demonstration in the first place. Even more so,
members of action groups, political parties or socio-economic classes seem
to be the people who typically might be expected to share an ideology. For
our cognitive analysis here, thus, it is sufficient to know that ideologies are
shared (as well as acquired and used) by social groups or collectivities.
Although ideologies are a property of social groups, individual members
may of course 'have' or 'participate in' an ideology as group members. That
is, they may personally endorse, accept or use a group ideology in their
everyday practices. In Chis respect, ideologies are like natural languages.
Languages such as English, Chinese or Kiswahili are also (knowledge)
systems that are essentially social and shared by the members of a group —
the speakers of those languages. But that does not mean that the members of
such a speech community do not know and use the language individually. In
a similar way, I shall say that ideologies are to be defined as ideologies-ofgroups that may be individually (and as we shall see, variably) used by the
members of the group. This way of formulating the shared nature of
ideologies emphasizes the group-based, societal dimension of ideologies,
while at the same time accounting for the role of ideologies in the (variable)
practices of social members in everyday life. Both theoretically and empirically, Chis relationship is crucial, since we are able to actually observe
ideologies 'at work' only in diese social practices, as is also the case for the
manifestations of language systems or grammars
It should also be stressed that as soon as we talk about ideologies or other
beliefs as being socially shared, Chis involves a mode of generalization and
abstraction. This does not mean that, as individuals, social members all have
identical copies of such beliefs and ideologies. Rather, it will be assumed
that each member may have a personal version of the shared belief or
ideology, a version that is obviously a function of individual socialization or
ideological development. Some people may only have a rudimentary (and
perhaps rather incoherent) personal version of the ideology, whereas others
('the ideologues') have a much more detailed and consistent one. This
notion of personal versions of ideologies also accounts for the frequently
found individual differences (and even contradictions) in the expression of
ideologies in empirical research. 8 This does of course not imply that
therefore there are no shared, social beliefs or ideologies, no more than that
individually variable knowledge and uses of language implies that there are
no grammars. The point is only that as soon as we talk about groups and
their knowledge or ideologies, we abstract from such individual differences.
The distinction between personal and social beliefs is handy for many
cognitive and social reasons. Probably the most compelling reason to make
this distinction is that social beliefs may be assumed by group members to
be known to most of the other group members. In discourse this means that
social beliefs may be presupposed by the speaker, and need not be explicitly
Social beliefs
31
asserted as new information. In that respect discourse is like the proverbial
tip of the iceberg most of its implied or presupposed meanings remain
(mentally speaking). Many of the facts of everyday life are thus
routinely presupposed in talk and text, for instance that Bosnia is a country
in former Yugoslavia, what genocides are, what civil war is, and so on.
Cultures are typically characterized (also) by such bodies of shared beliefs.
As we shall see in detall later, the same is true for the shared and taken-forgranted (often commonsense) beliefs that define the ideology of a group.
Socio-culturally, shared beliefs have a number of further characteristics.
For instance, as briefly suggested aboye, most of these beliefs have a general
or abstract nature. That is, they are mostly not about concrete facts, but
about general properties of facts. If we know what a civil war is, we know
about civil wars and their properties in general, and we may apply such
knowledge when observing or speaking about all possible civil wars.
Distinctions between specific civil wars are thus abstracted from. In logical
terms, we may say that episodic knowledge consists of beliefs that can be
described by propositions that have constants, referring to particulars,
whereas social knowledge consists of beliefs that may be represented with
propositions with variables. This is not surprising, because the very fact that
beliefs are socia ly shared implies that they are used in many different
situations.
Particular versus general beliefs
At this point, however, the common distinction between personal/episodic
and social/general beliefs meets its first hitch. For instance, the civil war in
Bosnia is on the one hand a specific event (or a collection of events), but
knowledge about it is not merely personal, but widely shared, and hence
social, and at the same time not abstract or general. It is defined by a
particular location, time period, participants and actions. Because, as I did
aboye, we talk about 'the' civil war in Bosnia, the use of the definite article
presupposes that there actually is or was such a civil war, and that we know
that recipients know it. The question then is whether this kind of knowledge
is episodic (personal, particular) or rather social, or maybe both? Do we
need a further distinction in the 'system of beliefs' in memory?
Such a further distinction may indeed be useful. This means that both for
personal and social beliefs we may further distinguish between particular
(episodic, context-bound) and general (abstract, context-free) beliefs.
Thus, beliefs about the civil war in Bosnia would be an example of
particular social beliefs which may be shared and presupposed like any other
social knowledge of a more general or abstract kind, such as the knowledge
we have about civil wars in general. Another term that may be used to refer
to such shared social knowledge about particular people and events would be
listorical knowledge'. The important point is here to remind ourselves of
the fact that not all social beliefs are general, abstract or context-free. In the
same way as personal knowledge in episodic memory represents peoplé s
32
Cognition
personal experiences, we may thus say that historical knowledge is about
specific 'collective' experiences of a group, society or culture. The Holocaust may be prototypical of such a collective group experietice — and its
shared representation in social memory.
On the other hand, in my personal knowledge systems, I may have
knowledge about concrete personal experiences, such as the fact that my
friend Ruth went abroad yesterday, but also more general or abstract
knowledge of a personal nature, such as the fact that Ruth is a friend of
mine, that my neighbour is also a professor, that I always do my shopping on
Saturday morning, and so on. That is, I do have personal knowledge that is
not about concrete, particular events, but represents a more general state of
affairs (e.g. friendship), habitual events, or personal properties. The notion
of 'self may be defined in terms of this abstract, personal knowlédge. In all
these cases, this knowledge does not apply to unique events, actions or
situations, but to many instances of them in my personal life.
Such general personal beliefs may monitor my specific social practices in
a similar way to that of more general, socially shared beliefs. But they are
still personal knowledge, because I may not generally assume that most
other people in my group or culture share these beliefs with me. In a strict
sense, even when family members, friends or acquaintances (or when I am
famous, many others) may know some of these beliefs, my own personal
beliefs are strictly speaking individual: they define me as a unique person,
and their description would fill an autobiography.
Ideologies as general social beliefs
Having made these distinctions, ideologies may be assumed to be constituted by socially shared, general beliefs. That is, they do not feature beliefs
about specific, historical events. Our specific, historical knowledge and
opinions about the civil war in Bosnia may well be influenced by ideologies
(e.g. those of nationalism, pacifism, and so on), but they are not themselves
part of such a more general and abstract ideology. Similarly, although my
personal beliefs may also be influenced by ideologies, they are not socially
shared by a group and hence, as such, are not part of ideologies.
Again, we may compare this definition with that of language: my personal
language use is of course controlled by a socially shared grammar and tales
of discourse, but it is not properly part of such abstract knowledge of the
language. Of course, we may define a language empirically in terms of the
set of its actual manifestations in language use, but this is hardly the case for
the socially shared, abstract systems of mies of the grammar In that respect,
ideologies should rather be compared to grammars than to language defined
in terms of the infinite set of its 'uses'.
For the moment, we shall ignore these personal beliefs and individual
uses' of ideologies, but later we need to show how they may be influenced
by social beliefs. This link between the social and the personal is crucial,
because most social practices, and hence most discourse, are by definition
Social beliefs
33
accomplished by one or a few individual persons in particular contexts. That
is, if we ever want to explain that social practices or discourses are
ideological, or that ideologies are reproduced by them, we need to establish
the theoretical relationships between the social and the personal, the general
and the particular, the group and its members, and the abstract system and its
specific instances or uses.
Knowledge and opinions
Having made a distinction between personal and social beliefs and their
corresponding domains or functions of memory, let us now further examine
the kinds of belief that define the social mind.
We have earlier' seen that ideologies are often assumed to tell groups and
their members what is good or bad, wrong or right. That is, ideologies
feature evaluative beliefs or opinions. More specifically, since ideologies by
definition are social and shared, they feature the social opinions of a group.
And because social beliefs are often general and abstract, so are these social
opinions, for iñstance the general opinions feminists have about gender
inequality. In fact, as we shall see in more detall below, the social opinions
that constitute an ideology are so general and abstract that they organize
clusters of domain-specific social opinions of a group, namely, attitudes.
Thus, general opinions about gender inequality in a feminist ideology may
be assumed to underlie a large number of more specific feminist actitudes
about, for example, discrimination and harassment on the job, unequal
political power, and so on.
Given the distinctions made aboye between social and personal beliefs,
we may further assume that this distinction also applies to opinions: besides
the social opinions we share with other group members, we also have
personal opinions, which are stored in episodic memory. We shall see later
that such personal opinions may of course be influenced by the social
opinions of the groups individuals identify with. Obviously, also these
personal opinions may be general (1 love ice-cream','I like my neighbour ^
or specific, that is, evaluations of specific personal experiences enjoyed
teaching in Valparaiso last monthD.
There are many other cognitive and discursive ways to characterize
opinions. One typical (though not exclusive) property is that opinions vary
contextually, or within a group or community. An opinion thus presupposes
that there are possible alternative opinions. It does not make sense to apply
truth criteria to a social opinion: 'We don't want any more immigrants' is a
xenophobic opinion that is neither true nor false, but a belief with which one
may agree or not, or that allows us to judge che opinion holder. Opinions are
typically entertained or expressed from a specific position or perspective, by
a person or a group, or in a specific situation, and are therefore also called
points of view. Opinions are not beliefs that tell us something about the
world, but rather about people who have them, or about the relations
(judgements) people have to the world.
34
Cognition
Attempts to define the notion of opinion, as we see, bring in another major
type of socially shared belief, namely, knowledge. Whereas opinions define
what we like or dislike, what is good or bad for us, or what must or should
not be done, knowledge is defined in tercos of what (we think) is the case,
what is true or false. Whereas opinions, as evaluative beliefs, presuppose a
judgement based on socially shared values and norms, our socio-cultural
knowledge consists of socially shared factual beliefs based on socially
acknowledged truth criteria. These truth criteria or rules of evidence may be
those of everyday common sense (dependable perception, reliable communication, or valid inference), those of science, those of religion, or any other
evaluation basis, depending on the social domain, group or culture for which
truth or factuality must be socially established.
Factual beliefs may be true or false. Thus, the proposition'The Hague is
the capital of the Netherlands' is a factual belief, although it happens to be
false. 9 It does not imply an evaluation, and its truth value can be established
by 'objective', generally accepted, truth criteria. When describing people
who entertain a factual belief that we think is false, we typically do so with
the verb 'believé : 'Larry believes that The Hague is the capital of the
Netherlands.' On the other hand, knowledge is true factual belief, and the
socio-cultural knowledge we deal with here consists of socially or culturally
shared factual beliefs that are true according to (similarly socio-culturally
shared) truth criteria. Similarly, at the interpersonal level, we say that
someone Inows' something if we think that what he or she (factually)
believes is true. In other words, both at the interpersonal and the social level
of analysis, knowledge is closely associated with sharing factual beliefs and
sharing the criteria for the establishment of the truth of such beliefs. This is
only a first approximation, and we shall come back to the definition of
knowledge and its relations to other beliefs (including ideologies).
The distinction between knowledge and opinion is very old, and goes
back to the opposition between, respectively, epistémé and doxa in classical
Greek. Relevant for Chis chapter is that we fmd it theoretically useful to
distinguish between socially shared evaluative beliefs or opinions (and
attitudes), on the one hand, and socially shared factual beliefs, or knowledge,
on the other hand. Whereas social opinions are based on values and hence on
the moral order of society, factual beliefs draw on what we may can its
'epistemic order', that is, the underlying system that features the basic truth
criteria for beliefs about the world.
This distinction is deeply embedded also in our commonsense thinking
and judgements about the world. Social members routinely distinguiste
between beliefs or statements about the objects or events of the world, and
those that involve their personal or social relations or positioning ('attitude)
with respect to the properties of these objects or events, for example as being
desirable or undesirable. For instance, people distinguish what they know
about abortion from what they think about it. They know that knowledge
may be culturally shared ('we all' know what abortion is), but that opinions
usually vary between different persons or groups ('we' are Pro-Choice, but
Social beliefs
35
'they' are Pro-Life in the abortion debate). They know that, in discourse,
cultural knowledge is often presupposed, whereas opinions usually need to
be defended.
Despite these (and other, more formal and philosophical) criteria, the
distinction between knowledge and opinions is very hard to make explicit.
What for some people or in some contexts is called'knowledgé, may be an
'opinion' for othérs or in other contexts. We might say that opinions are
represented by propositions that feature evaluative predicates, that is,
predicates that presuppose values, whereas factual beliefs do not. Por many
examples such a criterion would apply nicely, but for other cases the
distinction between evaluative and non-evaluative predicates is not that
clear.
Thus, the beiief'Amsterdam is the capital of the Netherlands' is obviously factual, and 'Amsterdam is a beautiful city' is clearly an opinion. But
that does not mean that'being the capital of cannot be used evaluatively, as
in the accusation'Amsterdam is the capital of drugs.' Similarly, we may use
the apparently factual predicate 'is a village' as part of an evaluative
proposition, as in 'Amsterdam is only a village, when compared with New
York.' We may factually conclude from a verdict that Henry is (was
convicted as) a thief, and still have no opinion about Henry, but at the same
time entertain the opinion that Harry is a thief because he stole some of my
ideas. Many predicates may thus have a more factual or descriptive and a
more evaluative meaning or use, as is the case for 'big', leavy' or
'clangerous'.
The same is true for socially shared beliefs, and hence for the distinction
between socio-cultural knowledge and social opinions or attitudes. Thus,
smoking may generally be found to be 'dangerous for your health', and the
belief that this is the case may be qualified as a factual social belief, which
can be proven to be true by generally accepted truth criteria, such. as
scientific experiments or statistical evidence, for instance as established by
the Attomey General in the USA. At the same time, there are variable social
opinions about smoking, even about its 'alleged' danger. Feminists will
claim that gender inequality is a fact, and adduce statistics to prove it,
whereas many conservative men (and some women) may disagree. In other
words, at some level of analysis, the distinction between social knowledge
and social opinion is not that clear. It is at this point also that ideology may
be involved in the distinction.
Epistemological approaches
Also the intricate accounts of knowledge and beliefs in contemporary
epistemology offer little assistance in establishing clear-cut theoretical
criteria for the distinction between knowledge and opinions. The ingenuity
of theorists as well as ordinary language users nearly always provides
counter-examples to most formal accounts. Thus, knowledge (of a person A)
is traditionally defined in tercos of the conditions (a) p is true, and (b) A
36
Cognition
believes that p. But this aliows for the possibility that A may just have a
lucky guess (as during a multiple choice test), and correctly believe that p, so
that we also need to add a third condition (c) A is justified in believing p.
Such justification should be based on the truth criteria mentioned aboye.
One problem with such abstract philosophical definitions is that they tend
to ignore the social and discursive contexts of truth conditions, criteria and
justification. That is, the actual use of a statement type like 'Chandra knows
that p' does not presuppose that p is true, but that the speaker (also) believes
thatp, and believes that there is enough evidence for p. That is, the problem
of the conditions of Chandra' s knowledge reverts back to the problem of the
knowledge of the speaker, so that we are back at square A. This means that
social issues of intersubjectivity and consensus become involved here. The
same is true for the acceptance of the truth criteria by which someone is
thought to be justified in her or bis beliefs, criterio that are historically and
culturally variable. In, our contemporary culture, such criteria may ultimately
be those of 'sciencé , but there are also known to offer no ultimate
'foundatioñ . In sum, somehow always social and cultural criteria of
knowledge (and hence of opinion) become part of a more empirically
warranted account of knowledge and beliefs. Abstraction from such social
contexts, and trying to find a context-free definition of knowledge, thus
seems to create more problems than it solves.
Hence, in this tripartite cognitive—social--discursive approach, we do not
deal with'abstract' knowledge, but with mundane talle and thought about
real personal or social knowledge, according to which A is said to'knowp'
if A believes p and also the speaker, or a whole community, believe that p.
This of course makes knowledge relative, but there is no way to escape such
relativism. True, the speaker and the whole knowledge-community may be
in error aboutp (and there are many historical instances where this was the
case), but in order to decide that such is the case another speaker-knower
(from outside the community) needs to establish this error in the first place,
so that knowledge is again made relative to that speaker-knower, and so on.
In other words, for beliefs of people to be promoted to the status of (true)
knowledge, we have no practical or theoretical means to escape the
consensus of some belief community by whose criterio A' s beliefs are
deemed to be true. Moreover, such a philosophical approach does not offer
an account of the difference between factual and evaluative beliefs — by
which criteria other than social ones are we able to establish that 'Henry is a
thief is true?
Cultural versus group beliefs
Therefore, in order to solve some of the theoretical puzzles of the distinction
between knowledge and opinion, let me make a further distinction, a
distinction we also will need in the further definition of ideology, namely,
that between cultural beliefs (or societal, or simply 'commoñ beliefs) and
Social beliefs
37
group beliefs. Although both the notion of 'culture' and that of 'group' are
themselves fundamentally ill-defined (see Chapter 15), the point of the
distinction is to differentiate between general, taken-for-granted beliefs of a
whole society or culture, and the more specific, often partisan, beliefs of
various social groups within such an oyeran culture. As we shall see in more
detall below, ideologies typically belong to the second type of beliefs. We
shall assume that ideologies form the foundation of such group beliefs.
One of the more specific reasons why we need this distinction is the
following. As suggested aboye, some groups in society have beliefs which
they qualify as knowledge, whereas others (other group members) qualify
these beliefs either as falsé factual beliefs or simply as opinions. If such is
the case, the very epistemological as well as cognitive theory of knowledge
becomes a prétarious undertaking. We would need to assume a relativist
theory of knowledge, according to which all knowledge is relative to a group
or culture. Any belief 'we' (members of our group) would hold to be true,
and which is shared by everyone in our group, could in principie be
challenged by others as false or as an opinion.
We shall further explore these relations between beliefs, knowledge and
opinions in Chapter 11, and here only make some overall assumptions about
the furniture and architecture of the social mind as a theoretical construct. In
earlier versions of this theory, it was postulated that ideologies are the basis
of the social mind. Although this explains how ideologies organize people' s
attitudes, this would also predict that all knowledge, as one major part of
social cognition, is ideologically controlled. Whereas this is undoubtedly the
case for many types of knowledge, especially knowledge about the social
world and knowledge that involves different interests or goals, this is not a
very plausible assumption. Moreover, if all our knowledge is ideological, the
notion of ideology loses much of its explanatory power. People have vast
amounts of everyday, commonsense knowledge about the world that neither
seems to be contested nor is obviously ideological. How then should a basic
ideological system control or organize some parts of our socio-cultural
knowledge and not other parts?
I therefore decided to put the original architecture on its head, or, if you
like, back on its feet. Instead of defining ideologies as the basis of all social
cognition, we will now assume that general, cultural knowledge is the basis
of all group-specific beliefs, including ideologies. Such cultural knowledge,
or cultural common-ground, may be defined as the (fuzzy) set of those
beliefs that are shared by (virtually) all competent members of a culture, and
that are held to be true by those members by similarly shared criteria of
truth. This is also why we may simply call this the repertory of 'common
knowledge' of a culture. It is this knowledge that all new members of a
culture have to leam (e.g. during socialization, formal education, through the
media, etc.) in order to become competent members. As suggested aboye,
this is the kind of knowledge that in most social situations — for example, in
interaction and discourse — may be presupposed, and which is called
'knowledge' by all members. This knowledge consists of all uncontested,
38
Cognition
commonsense beliefs, as well as of those specialized (e.g. scientific) beliefs
that have been 'adopted by the culture' as a whole, for instance our
knowledge that the earth is round and not fiat and turras around the sun
(despite our everyday perception to the contrary). It should be emphasized
that this notion of cultural knowledge refers to a collective, social phenomenon. It says something about beliefs shared by a cultural community, and
not about the knowledge of all individual members. Children, the mentally
disabled, cultural newcomers and others who are not (yet) fully competent
may only pardy share this cultural knowledge. That is, foil cultural competence of each member may be measured by the amount of cultural knowledge acquired, at least passively or implicitly (not all knowledge may always
be accessible actively).
Contrary to this kind of cultural knowledge, different groups may have
beliefs that for them also constitute uncontested knowledge, in the same way
as cultural knowledge is accepted by the whole cultural community This
group-knowledge may be verified by truth criteria that are either generally
cultural, but differently applied, or by group-specific criteria. Typical
examples are the kinds of knowledge as they are generally accepted within
the sciences, the professions, religions or political groups. Interestingly,
most of these knowledges build on general cultural knowledge, because
otherwise inter-group understanding, communication and interaction would
hardly be possible. Some knowledge partly extends or substitutes commonsense cultural knowledge, as is typically the case for scientific, technical
or professional knowledge. In these cases also the truth criteria for verification may be much stricter or elaborate. Conversely, religions may share
knowledge ré.g. about God) and may adopt truth criteria (such as faith) that
are not shared outside the religious group. And finally, different sociopolitical groups may have specific knowledge about society and its groups
that is not (yet) common knowlédge — for instance feminist insights in
gender inequality, or ecological insights into forros of pollution. As suggested, some of these insights (whether scientific, religious or social) may be
adopted by the whole cultural community. Even some specialized truth
criteria of one group (such as statistical evidence, the use of specialized
machines, etc.) may become adopted as a truth criterion by the whole
cultural community And vice versa, what in one historical period was
culturally shared, common knowledge may be abolished by the cultural
community as a whole and later only be maintained by epistemically
'deviant' groups.
The distinction between cultural and group knowledge is recursive and
can be applied both to whole cultures, as well as to subcultures. That is, at a
historical, cross-cultural or universal level of description and explanation,
what is cultural knowledge for one culture, may appear to be specific groupknowledge at a higher level. Cultural conflicts as well as difficulties of crosscultural communication and interaction bear witness to this form of
relativity. The same reasoning would then leave open the possibility of a
stock of 'universal' knowledge, that is, beliefs that are shared by the
Social beliefs
39
competent members of all cultures. Again, many commonsense and everyday beliefs (about people, their bodies, about the weather, nature and basic
social relations) would be candidates for such a stock of everyday knowledge. It is hard to imagine cultures that would not have shared beliefs about
mothers and their children, about men versus women, about young versus
old people, about the parts of the body, about edible food, and so on.
Similarly, the culture versus group distinction also applies at lower levels,
that is, within cultures. Groups and their knowledge are often characterized
in terms of subcultures, within which specific groups may again be distinguished with their own knowledge system. Similarly, groups or subcultures need not be part of one culture, but may be constituted over cultural
boundaries, as is the case for professionals, scholars and members of
different religións or political ideologies.
Neither cultural nor group knowledge is a well-defined concept. They are
essentially fuzzy, in the sense that there is no effective procedure to establish
for each culture or group what beliefs they collectively share (or indeed
which are only shared by part of the group). Yet, the notions are far from
arbitrary, and a quite reliable test (and there are others) is presupposition in
discourse. Cultural knowledge may be presupposed in all types of discourse
by all competent (adult, sane, etc.) members, except of course in all didactic
and pedagogical discourse that serves to teach such knowle' dge. The same is
true for group knowledge, which may be presupposed by all group members
in all their discourses (except of course in didactic or initiation discourse, or
in discourses directed at other groups, such as propaganda).
It will be assumed that general cultural knowledge (whatever its further
structures, functions, acquisition and change) must be the foundation of
social cognition. Ah specific group beliefs as well as the very interaction,
communication and mutual understanding of members of different groups
presuppose such cultural knowledge.
Cultural knowledge is therefore also the basis of all evaluative beliefs,
including socially shared opinions, attitudes and ideologies, as we shall see
in more detall below. For instance, different groups may have different
opinions about abortion, nuclear energy or state control over the market, but
such different opinions presuppose general (as well as specific group)
knowledge about what abortion, nuclear energy, the state or the market is.
Prejudices against, say, Turks thus presuppose that we know at least that
Turks are a people and not a brand of ice-cream or sportswear, although (as
some research shows) we may know very little else about Turks than that
they are a 'foreign' people. In other words, opinion differences still need a
common ground consisting of a cultural basis of knowledge.
The concept of cultural common ground is most obvious for shared
knowledge. However, we may wonder whether it also applies to other kinds
of beliefs, such as opinions. Since opinions, nearly by definition, are the
kinds of beliefs people may disagree about, this does not seem very likely at
first sight. Yet, in the same way as we have a general epistemic order, there
may be a culturally shared moral order, featuring the uncontested opinions —
40
Cognition
as well as the principies of moral judgement, that is, cultural values — of a
given culture. Just as specific societies have laws as well as a constitution,
thus, cultures have a moral basis which again monitors interaction, communication and discourse across group boundaries. Again, such moral principies should be uncontested and presupposed in all evaluative talk, action
and interaction. They are also the basis for judgements about and sanctions
against moral deviance by individual members of a culture.
In the same way as specific group knowledge presupposes cultural
knowledge, group opinions and their underlying norms and values should
presuppose the culturally shared moral order. And the same top-down or
bottom-up dynamic may be at work here. What the norme, values and
opinions of a specific group are may gradually become culturally shared by
a whole culture, and vice versa. What once was a culturally shared norm or
opinion may later become characteristic for a specific group. For instance,
whereas Christian religion may once have been constitutive of the moral
order of much of Western culture, it has now been reduced to that of a
specific religious group. And the basic normative system of human rights
which once was specific to groups of philosophers in the eighteenth century
is now largely accepted throughout Western (and other) cultures.
The oyeran architecture of the social mind we have now constructed is
one which has a general cultural basis of common factual and evaluative
beliefs. This cultural common ground is acquired and accepted by virtually
all members, and presupposed in all discourse and other interaction. It is on
this basis that different groups may develop specific knowledge and opinions, and may compete for epistemic or doxastic hegemony, or indeed even
for (pardal) acceptance in the general common ground for the culture(s) in
which they participate.
The same is specifically true for ideological competition and struggle.
Whereas it was earlier assumed that ideologies are the basis of social beliefs,
and we added that these were the social beliefs of a specific group, we meant
just that — ideologies will be defined as the basis of social group cognition.
In this case, it is perfectly acceptable to have them control both the opinions
or attitudes of the group, as well as their knowledge, because specific group
knowledge may very well be related to the interests or other properties of the
group, and be involved in competition, struggle or domination.
This way of organizing the social mind also implies that as soon as social
beliefs enter the set of general cultural beliefs, they by definition are no
longer ideological for that culture, but simply basic knowledge or opinions
that are shared by everyone, taken for granted, and uncontested. Of course,
another culture (or the same culture in a later period) may of course again
deem such beliefs to be ideological. In other words,Ideó1.51"2 - always
presuppose specificity for a group or culture, and hence competition,
confrontation, or at least evaluative comparison at a higher level or from a
point of view outside the group or the culture.
This also elegantly solves the relativity problem for knowledge and other
beliefs. If we assume that there is no absolute knowledge, and hence no
Social beliefs
41
ultimate truth criterio, we still do not need to be relativist with respect to a
given culture — knowledge may well be accepted as true within a given
culture, given the truth criteria of that culture. This may even be the case
within each group, whose members will claim that their beliefs are trae,
whereas those of other beliefs are false factual beliefs or merely evaluative
opinions. When dealing with the relation between knowledge and ideology
in more detall (Chapter 11), I shall elaborate this point.
Types of beliefs
Before I continue my analysis of the contents and organization of the social
mind, let me recapitulate the kinds of social beliefs and distinctions we have
encountered so faz:
• personal versus socially shared beliefs
• specific versus general or abstract beliefs
• specific social beliefs or historical beliefs
• factual versus evaluative beliefs (opinions, attitudes)
• truth criteria versus evaluation criteria (norms, values)
• true factual beliefs (knowledge) versus false factual beliefs (errors,
illusions)
• cultural beliefs (common ground) versus group beliefs.
These distinctions also imply, thus, that beliefs in general should be
described as group beliefs (G-beliefs) and cultural beliefs (C-beliefs), and
the same is true for knowledge and opinions. Usually, when we speak about
knowledge, we mean C-knowledge, and not G-knowledge. The latter type of
knowledge is accepted only by one or several groups and is often simply
called 'beliefs' (e.g. 'They believe that God exists', They believe that the
market will solve all social problems), or opinions, illusions, myths, fiction,
fallacies, and so on, by other groups. Ideologies, as we shall see, are the
general, social beliefs that are the basis of G-beliefs. And cultural beliefs
form the common ground of (virtually) all social beliefs of (virtually) all
groups of a given culture.
At the same time, these distinctions provide a framework for the social
dimension of the classical opposition between objective and subjective (or
intersubjective) knowledge and beliefs. If objective knowledge consists of
those beliefs that are shared by everyone, and can be shown to be true by
the truth criteria of a community, then such objectivity may also be Cobjectivity or G-objectivity, depending on whether they are shared by one or
more groups or the whole culture. As with knowledge, when we speak about
objectivity, we usually mean C-objectivity. Subjective beliefs are all those
beliefs that are associated with a specific person, group or culture, and which
are not accepted by all members, all groups or all cultures, respectively,
depending on the perspective or scope of the description.
These distinctions are not merely the fruit of cognitive or philosophical
speculation, but rather specific hypotheses about the organization of memory
42
Cognition
in general, and social memory in particular. They are necessary in order to
define ideology, and to solve the well-known problem of the differences
between knowledge and ideology. Moreover, they are used in describing and
explaining different discourse structures. Knowledge and opinions are
expressed and sustained in different ways in discourse', and require different
forms of 'evidencé
We have also seen that in natural language and commonsense discourse
the notions 'belier,, 'knowledge' and 'opinion' may be used differently than
we did aboye. On the other hand, we have tried to make explicit some of the
implications of the everyday uses of such terms. Instead of the cognitive
distinctions made aboye for social beliefs, we may make similar distinctions
in the discourses that express or construct such social beliefs. Instead of
'beliefs', we might then account for different types of discourse in terms of
the different types of descriptions they give of the social world."
As we have argued before, there are many reasons why we do not adopt
this kind of discursivist reduction. In this case, for instance, although the
cognitive distinctions should be shown to be relevant for discourse description, language users are not always able to make explicit their knowledge of
different forms of social beliefs. More generally, then, it is important to
.
10
distinguish carefully between beliefs and the expression of beliefs in dis-
course. The latter are also a function of the constraints of the context,
including personal beliefs or experiences, and not only of the underlying
structure of social memory.
It has been assumed aboye that factual beliefs may be said to be true or
false, as are the propositions that represent them. However, it might be
argued that truth values only apply to actual statements or expressions of
beliefs, that is, in discourse, and only in the pragmatic context of assertions.
'Is The Hague the capital of the Netherlands? expresses a factual belief, but
there is no point in calling this belief 'true' or 'false' — indeed, the question
presupposes that the speaker does not know whether the belief 'that The
Hague is the capital of the Netherlands' is true or false. At most, the belief
may be said to be possibly true, given the presupposed knowledge that The
Hague is a city in the Netherlands, and the seat of the Dutch government and
parliament. Such a possibility may also be expressed by modal expressions
such as 'maybe', 'perhaps', and so on, which also express doubt about the
truth of a belief. In other words, factual beliefs are not just true or false, but
also possibly true or false. In formal terms, they are not just propositions, but
propositional functions, which may be turned into actual (true or false)
propositions in contextualized discourse, and if they are asserted. In sum, if
we continue to speak about true and false beliefs, the social mind may
feature factual beliefs of which the truth status is unknown. We shall come
back to the discourse manifestations of social cognition in Part III.
In sum, both personal and social opinions imply differences of opinion,
that is, my opinions (versus those of others), on the one hand, and our
group's opinions (versus those of other groups), on the other.
Social beliefs
43
Attitudes
I shall use the term attitude to denote general, socially shared, evaluative
beliefs (opinions) of groups. Or rather, I shall reserve the notion of attitude
to refer to specific, organized, clusters of socially shared beliefs, such as the
(often complex) attitudes about nuclear energy, abortion or immigration.
This means that, contrary to the sometimes confused usage in social
psychology, personal opinions are not called attitudes, whether or not they
are particular or general. Individuals may of course 'participate in' or share
a social attitude, as they also may share in social knowledge or know a
language. I shall for the moment leave open the possibility that the notion of
attitude may also apply to clusters of socially shared particular opinions, for
example about this civil war in Bosnia and not just about civil wars in
general. Although social opinions and hence attitudes typically vary between
groups, we might also speak of cultural actitudes if a cluster of social
opinions is shared by a whole culture, as might typically be the case for
cultores defined by one religion. 12
Why the 'attitude' concept cannot be dispensed with
Some social psychologists have criticized the traditional notion of attitude
on more fundamental, anti-cognitivist grounds. They dispute that people
'have' something like attitudes in the first place, and that such attitudes
control people' s actions or discourse. According to these critics, opinions or
attitudes do not 'exist' at all, at least not as 'fixed' mental representations.
They emphasize that opinions (like the mind in general) are social constructions. Moreover, these scholars emphasize that opinions should be defined in
terms of their discursive formulation. For them, opinions vary with the
context in which language-users rhetorically engage in debate or other
interaction with other participants. Instead of attitudes, discursive 'repertoires' are proposed to account for such variations in the formulation of
opinions. And if attitudes should 'exist' mentally at all, they should rather be
dynamically represented as some kind of rhetorical structure, or as an
argument. 13
As has been emphasized before, there are many arguments why this
position is theoretically untenable. A detailed discussion of this issue is
beyond the scope of this book, so a few succinct arguments will have to
suffice here to reject this approach to attitudes.
(a) In more general terms, it has already been shown that a reformulation
of cognition in tercos of discourse is a form of interactionist (if not
behaviourist) reduction that fails to describe and to explain fundamental
properties of both thinking and discourse. If all'non-observablé mental
entities would need to be dispensed with, we would also have to throw out
beliefs in general, including knowledge, tules, and of course discourse
meaning, among many other cognitive notions. Moreover, interaction and
discourse structures thernselves are abstract and hence unobservable. The
44
Cognition
same is true for other practical and theoretical Unobservables', such as
groups, power, inequality, institutions, society and culture, which we also
postulate (in a social theory) in order to be able to describe and explain
people's activities ('behaviour'), among other things. In sum, if 'observability' were a criterion, neither commonsense nor theoretical analysis of
action, discourse or society would be possible, no more than an analysis of
people's minds.
(b) To abolish the 'mind' as a practica! and theoretical notion for
everyday and scholarly observation and explanation, and without providing
a serious alternative, is not only counterintuitive, but also inconsistent with
all available evidence. That minds are obviously (also) social constructs does
not mean that they do not 'exise , as a specific and complex property of
people's brains. An interactionist, discursivist or constructivist reduction of
the mind is unable to explain what people do when they think, believe, have
opinions, remember, and so on.
(c) Opinions (and hence attitudes as socially shared opinions) also
underlie other social practices than discourse, as is the case for prejudices in
relation to acts of discrimination To reduce prejudice to (say) verbal
'repertoires' (whatever these are exactly) is to deny that discrimination may
be 'based on' prejudice, or to deny that prejudices, for all practical, social
and theoretical purposes, 'exist' independently of discriminatory behaviour.
Moreover, such social opinions and attitudes are gradually acquired and may
change, and hence are not
although, at the level of the group, they
should remain relatively stable across several contexts of their application.
(d) Both for common sense, and theoretically, people or groups are
assumed to 'have' opinions and other beliefs also when they do not always
express them in talk, or in other social practices for that matter.
(e)That people usually tailor the precise formulation or expression of an
opinion to the constraints of different contexts, does not imply that the
underlying personal opinion itself may not, for all practical purposes, be the
same in different situations. We know that persuasion and rhetoric may fail,
and that people often'do not change their opinion'. This is especially the
case for socially shared opinions, which by definition can only be shared
when they Are not different from one local context to the next, and hence
from person to person.
(fl The reduction of opinions to their ad hoc formulation is inconsistent
with a basic condition of social interaction and social groups: that social
members may share a 'common ground' of beliefs. Paradoxically, radical
social constructionism that denies mental beliefs is inconsistent with its own
social claims, and reduces beliefs (and ideologies and culture) to the
solipsism of interacting individuals in unique contexts.
(g) Of course, personal opinions (whether shared with a group or not)
may — but need not adapt to specific social situations or contexts. But this
does not entail that therefore they are not mentally represented. As I shall
show later, people represent their personal and local knowledge and opinions
Social beliefs
45
about an event in mental models (see Chapter 7). It is this representation of
(personal) opinions in models that explains cóntextual variation, and also
provides a solid ground for the explication of both discourse and other social
practices in which such opinions are expressed. Mental model theory
elegantly explains all objections against the postulation of socially shared
attitudes, and does not have the numerous problems inherent in reductive
'discursivism'. Thus mental models allow for shared social opinions or
attitudes to be relatively stable (although they may change in time), while at
the same time providing for individual and contextual variation and
uniqueness.
(h) And finally, alternative proposals, such as 'repertoires' or 'rhetorical
mental structures' are either left undefined as to their precise structure and
status, or in fact alsó boil down to something (unobservable!) people !have'.
They are a form of knowledge or belief, and hence mental. After all, we can
hardly assume that repertoires are floating in the air or in people' s mouths. If
they allow people to talk or understand talk and text, we have no alternative
but to locate them in the minds of people, as is the case for grammars,
discourse mies, norms, and indeed knowledge and other beliefs.
The critique of the classical notion of attitude is correct in concluding that
(besides many other flaws) traditional social psychology largely ignored the
crucially discursive and social nature of attitude construction and manifestations, and underplayed the contextual variation of attitude expression.
However, this is no reason to throw out the baby with the bath water, to deny
that opinions and attitudes do not 'exise , to claim that they are merely
mentalist 'reifications' and that where relevant they exist only as discursive
formulations.
Denying the existence of attitudes because they are 'unobservable' would
in this case be as silly as affirming such existence, simply because there
would not be any direct evidence for either claim. This is the case for all
properties of the mind. They are being postulated, practically and theoretically, because they are real in their consequences: they explain how and
why people can 'meaningfully' and 'purposefully' act and talk. They explain
very powerful commonsense self-observations: people know they think, they
know they know things, and they know they 'have' opinions, whether or not
they express them, and even if they express them differently in different
situations. People know that often they agree with others, and may thus
share opinions as members of a group. The 'silent majority' is defined in
terms of such a community of people sharing the same or similar attitudes,
even if they do not always express them. There is no more 'reification'
involved here than in the commonsense and theoretical assumption that
people have contextually variable knowledge as well as more general sociocultural knowledge that may be (variably) used in different contexts.
In sum, in a more explicit theoretical framework which describes their
precise status, their interna! organization, their cognitive and social functions, 'attitudes' remain a useful concept. To reduce opinion clusters on, for
46
Cognition
example, immigration or nuclear energy, to forros or repertoires of talk, is to
confuse levels of description and explanation, to ignore that manifestations
of human activities may have explanatory nhderlying structures, and to
challenge commonsense observations, without providing a theoretically
sound alternative. It is like saying that feeling hungry does not 'exist'
because we can' t see that, and that such a feeling should in fact be described
only in terms of people ingesting lots of food. We know that people are
hungry (also when they are not eating) because they are able to tell us, in the
same way as they are able to tell us that they feel bad about being hungry, or
that they have the opinion that poverty is due to the riches of the rich.
Opinions, thus, are not less real than hunger, and should not be reduced to
their manifestations in discourse or social practices.
Concluding this brief and incomplete argument, we find no grounds to
abolish the notion of attitude. On the contrary, especially also in a theory of
ideology, such a notion, when properly analysed, is crucial. It accounts for
the 'common ground' of socially shared opinions of groups of people and
for the ways these allow group members to interact, to coordinate and to
organize their social practices, even in different contexts. What we do need,
however, and what was another major shortcoming of much traditional
attitude research, is a much more detailed analysis of their interna organization. Similarly, we need to examine in more detall how socially shared
opinions or attitudes are linked with personal ones and in different contexts.
And of course, we should spell out the social situations and social structures
in which social groups develop and change their attitudes, and especially
how they do so exactly. And finally, yes, we must account for the ways both
social and personal, general and specific opinions are being expressed and
formulated in text and talk. These are some of the tasks of the next
chapters.
Social representations
So far, I have used the general notion of belief in order to describe personal
versus social beliefs, specific versus general beliefs, factual versus evaluative beliefs and group versus cultural beliefs. As systems of knowledge and
attitudes, these beliefs are organized in many ways, for instance by schemalike structures such as scripts, scenarios, frames or other organizational
pattems of memory. In order to have a general concept that specifically
applies to organized clusters of socially shared beliefs (knowledge, attitudes,
ideologies, etc.) as located in social memory, I shall henceforth use the term
social representation, of which social beliefs are the constituent elements.
The concept of 'social representation' has been used in social psychology
and the other social sciences in many different ways. 14 Here, however, the
term 'social representations' (SR) will only apply to organized clusters of
socially shared beliefs. Thus, knowledge scripts and attitudes are both
examples of social representations, and so are ideologies. The next chapter
Social beliefs
47
will deal with the intemal structure and further properties of these social
representations. One crucial dimension of a cognitive approach is not only to
describe the structures of mental representations, but especially also the
processes or strategies of their social acquisition, use and change.
Habitus
Especially in sociology, another term used to denote socially shared representations is habitus, usually defined in terms of 'structured dispositions' for
social practices that are partly autonomous and partly a function of societal
structures." As I have done with the notion of ideology, habitus is
sometimes compared with a generative grammar in order to emphasize the
creative, active use social actors make of such dispositions. I shall not
further use this notion, because cognitively it is only very loosely defined. ft'
certainly is less explicit than the notion of (a system of) social representations, or social cognition, used in this chapter. Moreover, the concept of
'disposition' in the definition of this concept is psychologically inadequate,
if not circular, because it defines cognitive structures in terms of their
'output' (such as social practices) which precisely need to be explained in
terms of other, cognitive representations. Por instance, prejudice as a social
habitus should not be described as a 'tendency to discriminaté , but be
analysed in terms of mental structures in such a way that discrimination,
verbal derogation, disclaimers (We are not racist, but. .'), as well as many
other manifestations of prejudice can be explained.
Social cognition
I shall henceforth use the term social cognition to refer to the combination of
socially shared mental representations and these processes of their use in
social contexts. This usage is different from one of the uses of the terco
'social cognition' in current social psychology, where it often refers to the
more individualistic, information-processing approach to social memory
prevalent in the USA, as distinct from (mostly) European approaches to
social representations, social identity, social categorization and intergroup
relations.
In this book, I advocate an integration of these (and other) approaches to
social cognition. That is, on the one hand, it should be recognized that the
mental representations and processes of social beliefs and actions need to be
described in explicit detall, whereas on the other hand social cognition, and
especially ideology, can fully be understood only in terms of their social
functions for social actors as group members in social situations."
That much current work on social memory representations and processes
largely uses the prevalent information-processing metaphor of cognitive
psychology is no problem as long as we know it is merely a metaphor, and
as long as detailed processing theories provide insights that alternative
approaches do not provide. Also, as suggested, the use of this metaphor does
48
Cognition
not at all commit us to an individualistic approach to the human mind, as
long as we know that the' mind is socially constituted and used, and hence
mental representations should also be described in terms of their functions
for group members and whole groups.
Ideologies: a cognitive definition
The ideas developed aboye, based partly on current psychology, partly
extending them, provide the conceptual instruments for a provisional cognitive description of the nature and status of ideologies. In the following
chapter I shall add further details, especially also about the social dimen
sions of ideologies. In other words, we now only pretend to sketch part of
the overall theoretical picture.
We have discovered, aboye, that ideologies cannot simply be called
'belief systems', because there are many types of belief that are not
ideological in the usual sense, nor in the sense we would like to reserve for
the concept of 'ideology'. We need to locate ideologies in the social mind,
because they are not individual, contextualized, ad hoc beliefs, but socially
shared by collectivities of some kind Finally, we identified those socially
shared beliefs that need to be kept outside the control of specific ideologies,
namely, all culturally shared social beliefs, including especially the epistemic common ground of a culture.
The closest we got to the notion of ideology, thus, was to define them in
terms of the social beliefs shared by specific social collectivities or 'groups',
where the notion of 'group' needs to be defined later. This would mean that
an ideology is the set of factual and evaluative beliefs — that is, the
knowledge and the opinions — of a group. Depending on how we define
groups later, this is indeed quite close to the notion that is used most often in
commonsense and scholarly approaches to ideology, as we have seen
before.
This means that this chapter has given a first answer to the basic question
about the 'nature' of ideologies: They are not metaphysical or otherwise
vaguely localized systems 'of or 'in' society or groups or classes, but a
specific type of (basic) mental representations shared by the members of
groups, and hence firmly located in the minds of people. Thus, ideologies are
not 'aboye' or betweeñ people, groups or society, but part of the minds of
its members. Again, this does not mean that they are therefore individual or
only mental. On the contrary, just like languages, ideologies are as much
social as they are mental. It is this integrated socio-cognitive analysis that
characterizes my approach to ideology. In a more social and critical analysis,
I shall later have to examine the social, political and cultural conditions,
consequences and functions of ideologies thus defined — for example, in
terms of the values, identities, relations, aims, positions and power of social
collectivities of specific kinds.
Social beliefs
49
Ideologies as thefoundation ofgroup beliefs
I shall, however, further limit the concept of ideology in order to make it
even more specific as a theoretically viable notion and suggest that ideologies are thefoundation of the social beliefs shared by a social group. In other
words, a bit like the axioms of a formal system, ideologies consist of those
general and abstract social beliefs, shared by a group, that control or
organize the more specific knowledge and opinions (attitudes) of a group.
Forrnally, this would mean that the propositions that constitute an ideology
1 should be derivable from the variable knowledge and opinions about various
domains of social life. For instance, if ethnic prejudices pertain to human
rights, immigration, integration, education, housing, access to resources, and
so on, of minorities or immigrants, then the ideological beliefs would be
formed by such general propositions as 'We are fundamentally different
from them', We are superior to them', 'They are a threat to us', They do
not respect our norms and principies', We are tolerant', and so on. Later we
shall see how such propositions are organized in ideological schemas.
As may be expected, such basic ideological beliefs must be both general
and abstract, and also very relevant for a group. They typically would not
deal with details of everyday social life, but apply to fundamental dimensions of the group and its relations to other groups. As we shall see later,
they must be functional for the group as a whole, and reflect the conditions
of its existence and reproduction.
In an earlier version of my theory, I limited ideologies to the foundations
of evaluative beliefs. 17 The reason for this decision was that ideologies
generally apply to what is most characteristic of a group, namely, its
opinions about itself and other groups. However, once we have relegated all
forms of commonsense and general knowledge to the cultural basis of the
social mind, ideologies may also be taken as the basis of group knowledge.
This would mean that they do not only embody the specific values but also
the truth criteria of a group. For instance, Christians share the basic
ideological belief that God exists. Feminists typically as sume that women do
not have an equal share in society's resources. And ecologista have basic
knowledge about pollution and the relations between humans and nature.
Some of these general beliefs that originally were characteristic of special
groups, with 'special interests', have become part of the general cultural
common ground.
That ideologies control group opinions or actitudes, seems obvious.
Shared opinions must be important for the interaction, coordination and
reproduction of the group, and these judgements require values and general
principies that are typically variable across groups. They define competition,
struggle and inequality. Is this also true for specific group knowledge?
Provisionally 1 shall assume that this is the case: if factual beliefs are shared
by a group, then they are socially relevant enough to get an ideological
foundation. Also, the truth criteria in that case should be group specific,
because otherwise che beliefs would probably be part of che common cultural
50
Cognition
ground. Thus, protedures of proof, evidence and acceptance of beliefs in
scientific discourse and communication, are very different from those of
politics, religion, corporate business or the mass media, or indeed from those
in everyday life.
However, even within this much more specific scope, it may seem odd to
call all specific group knowledge ideologically based. This may be true for
religions, or even for action groups, where religious and political ideologies
determine how the world is understood, and where group interests may be
involved. But what about, say, the medical knowledge of doctors, the legal
knowledge of lawyers, or of our own scholarly knowledge? In some cases
there may not even be any competition, conflict or struggle from outside the
group. Yet, professional knowledge, as many studies and everyday experiences show, is a symbolic resource for professional, elite power. It is a
resource that is carefully protected, and serves the interests of the group.
Hence, it is plausible that the nature of that knowledge itself, and the ways it
is acquired, changed, validated and used, is also profoundly ideological.
Thus, medical knowledge of the body, which may seem scientifically 'true',
not only competes with religious and commonsense knowledges, but also
embodies the typical truth criteria and other principies and hence the
ideology (or ideologies) of the medical profession. Such knowledge may be
used and abused, it may be applied in order to control people, and it most
certainly is a fundamental condition for the reproduction of the profession. 18
Thus, although perhaps less relevant for each detailed piece of professional
knowledge, it seems likely that also professional knowledge as a whole is
internally and functionally characterized by underlying ideological principies.
It has already been suggested that as soon as the basic evaluation criteria
of a group, and hence its social beliefs, are increasingly adopted by society
as a whole, the group specificity of such social beliefs is lost, so that they are
no longer ideological in our strict sense, but simply part of the cultural
common ground. Of course, this does not mean that such a cultural common
ground itself may not be declared at an universal lev-el of
description and evaluation. As is increasingly clear in the contemporary
world, also whole cultures may clash, compete and hence have interests, so
that their shared common ground and its basis principies of evaluation may
again be ideological in comparison with those of other cultures. In other
words, if knowledge and other social beliefs are relative, so are ideologies.
If a general culture consists of generally accepted, uncontested beliefs,
which in fact define the shared common sense of its members, then we might
be tempted to call precisely such beliefs ideological. Ideologies have often
been declared really influential if nobody notices them, and if they define
common sense. This may be true, but it is inconsistent with the theory that
links ideologies with groups, group interests, group relations, struggle,
domination or specific world views. That is, we are only able to understand
and analyse common cultural ground as ideological if we have possible
alternatives, other examples, other cultures, conflicts between cultures, or
Social beliefs
51
when a specific group within a society or culture challenges the social
beliefs of the common ground. In other words, again, the relativity principie
applies: common cultural ground can only be called ideological at a higher,
comparative, universal or historical level of analysis. If all members of a
culture believe, for example, in the existence of God, then such a religious
belief is no longer ideological, but simply shared knowledge, within that
culture. That is, there are no groups within that culture that disagree, contest
or otherwise provide an altemative view of society in that respect.
On the other hand, if specific common-ground beliefs in fact are in the
interest of a specific dominant group (e.g. beliefs about the properties and
roles of women until not too long ago) and yet taken for granted, tacitly
accepted and uncontested by the other groups, then we already distinguish
between different grClups (e.g. men and women) and their different interests,
so that in that case such common-ground beliefs would in fact be beliefs that
are those of one group, as imposed on society or culture as a whole.
This suggests that parts of common ground may be ideological anyway,
but again this is true only with respect to a comparative or higher level, in
which we are able to distinguish different groups and conflicting interests in
such a society or culture. From within a totally homogeneous culture, no
conflicts of interests of any common-ground beliefs can even be perceived or
thought. As soon as one social group realizes that the common ground is in
fact not in the interests of all, then a set of common beliefs will be declared
ideological, and attached to a specific dominant group (e.g. whites, men). In
the same way as group beliefs may become cultural beliefs in many ways
(usually by power, hegemnny, inculcation, and so on), also the reverse may
be true, when individuals form a group that challenges generally accepted
social beliefs, develops opposed beliefs and hence its own ideology of
resistance.
These social constraints on ideology formation need to be attended to
later. It is, however, interesting that even within a cognitive account of
ideology, we need to postulate a social and cultural basis. In a social sense,
this requires social interaction, sharing, social situations, organization and
often also institutionalization. In the 'purely' cognitive sense (if there is such
a thing) taiking about the 'social mind' means first of all that cognitive
representations are not limited to individuals but in some sense distributed
among 'many minds' This presupposes information exchange — for example
through perception, discourse or interaction — which again brings in the
social dimension. Secondly, and more interestingly, we assume that the very
mental contents, architecture and organization of the social mind shared by
group members reflect social and cultural constraints. We are unable to
define 'knowledge' without having recourse to social or cultural conditions,
and the same is true for actitudes and ideologies. Thus, if we talk about a
cultural common ground of generally shared beliefs than this is not just a
socio-cultural account, but also tells us something about the very foundation
of the mind, of social memory and how other beliefs, including social ones,
are grounded and organized. Similarly, also the more specific group beliefs
52
Cognition
that members of different social groups develop, share and use, are differentiated only with respect to both this general common ground and to the
social beliefs of other groups.
We now have a first impression of the cognitive status and location' of
ideologies. The next crucial step is to examine what such ideologies actually
look like, how they are organized, and how they related to the social beliefs
(group knowledge and attitudes) of which they form the foundations.
0
Structures and Strategies
Modes of description
If there is anything a theory of ideology must provide, then, it is an account
of the structures of ideologies. Few topics in earlier approaches to ideology
have been ignored so consistently as the simple question: If there are
ideologies, then what do they look like? No sophisticated structuralism was
necessary to spell out the typical components, the building blocks of
ideologies, and how these are combined in various pattems. Yet, this seldom
happened, so that ideologies usually remained in an analytical limbo,
somewhere between 'systems of ideas' and 'social interests', where everybody could project into them what they wanted.
For contemporary psychology, linguistics and discourse analysis, as well
as for some of the social sciences, such questions of structure are routine —
describing, analysing and explaining phenomena first of all means that we
should specify their structures and their functions. Such analyses may be
static-structuralist, or dynamic proceSsual. The first, as we know it from
modem grammars, specifies the structural components or units, as well as
the principies (tales, norms or other regularities) of their composition in
larger units. 'The more dynamic approach, familiar in psychology, microsociology and conversation analysis, spells out the actual processes, moves
or strategies, that is, the mental or interactional dynamics of construction, for
instance as an account of how social actors or language users go about, online,'cloing or 'making' such structural units as mental representations,
actions or discourses.
Structural versus strategic analysis
I shall henceforth refer to these alternative modes of description as the
structural and the strategic approach. The first analyses objects as finished
products, the latter characterizes the ways in which such objects are
gradually built up or interpreted. These approaches may be seen as fundamentally different, as true alternatives, or as complementary ways of
accounting for the same phenomena, depending on oné s philosophy of
language, discourse, interaction or cognition. The more strategic approach
would then seem to account more adequately for what social actors, thinkers
or language users are actually doing in concrete situations, whereas the
structural approach would be more abstract and context-free, and rather
54
Cognition
account for what social actors know, or what the product or result is of their
strategic thinking or action. 1
At the moment, both in psychology, as well as in conversation analysis
and the social sciences, the more dynamic, strategic approach is more popular after the earlier, structuralist phase. Yet, as suggested, such approaches
are in fact complementary. First of all, both are abstract, both operate with
abstract categories, and both operate with some kind of rules. Even when
analysing the dynamics of cognitive processes or social interaction, we
operate at various levels of abstraction, with theoretical constructs accounting for what is being observed. Thus, conversation analysts may do so in
tercos of actions, turns, moves and their sequencing in talk, whereas
psychologists operate with cognitive units such as concepts, propositions,
mental representations or networks, and the strategies of their mental
manipulation in production and understanding. And neither cognitive
psychologists nor those who analyse interaction and conversation operate at
the various physical, physiological or auditory levels of 'reality'. That is,
any abstract account of construction processes or strategies presupposes
some kind of components or structural units known and used by information
processors as social actors.
That is, also a strategic approach assumes that speakers know what
structures are more or less well formed, and what rules or other structural
principies are available to them as (mental and social) resources when
engaged in strategic construction. In that respect, thus, the structural and the
strategic approach are complementary approaches to the description of the
various phenomena of cognition and interaction.
Similar remarks may be made for other, more complex social structures,
such as groups, organizations, group relationships, and whole societies,
which also may be structurally accounted for -in terms of their conceptual
building blocks and the principies of their construction, on the one hand, or
the strategic processes of their actual operation, construction, reproduction,
formation or change, on the other hand.
Abstract versus practical competence
There is, however, one difference between these two approaches that is more
fundamental. Structural approaches tend to be more abstract and contextfree, in the sense of characterizing ideal types or general patterns, and tend
to ignore variations, 'deviations' and 'errors'. Modem structural and generative grammars and earlier psycholinguistics usually take such an
approach. Under the influence of new directions in cognitive psychology,
socio-linguistics and conversation analysis, such abstract normativity was
relinquished for an account that focused on the on-line, ongoing processes or
strategies of what actors are actually thinking, saying or doing, including
individual, contextual variations and 'errors'.
Instead of the neatly separated levels of grammars or other structural
theories (e.g. those of argumentation and narration), and the theoretical
Structures and strategies
55
distinction between langue' and'parolé or between 'competence' and
'performance', the dynamic approach emphasizes that people think, speak
and act strategically. This means, among other things, that they follow
various goals, operate or act at various levels of production or understanding
at the same time and, while doing so, make mistakes, have lapses of
memory, get confused, or take short-cuts. Despite such'imperfectioñ, they
'are usually able to repair these and to re-interpret the data at hand. In sum,
they are clearly competent in managing this bewildering amount of tasks to
accomplish rather successfully, though not perfectly, what they set out to do,
namely, understand something, say something or do something in a specific
context, often jointly with other people. In that respect, the strategies of text
comprehension are not much different from those of conversation and
interaction. Both reqüire an abstract or normative competence, as well as a
more practical competence or ability.
Dynamic processes of thinking and acting are possible only when people
also know and share more abstract rules and structures. They often know
what sentences, sentence sequences, actions or interactions are more or less
well formed, acceptable or understandable. Such knowledge and judgements
are not merely displayed in ongoing discourse. Sometimes they may also be
applied in a more abstract, context-free fashion, because their knowledge is
not limited to one situation or to one token, but necessarily more general,
and hence abstract. This allows them to adequately produce and interpret a
potentially infinite number and variety of different perceptions, discourses or
actions. In sum, although the structural and strategic approaches have a
different flavour and focus on rather different aspects of thought, discourse
and interaction, they always presuppose each other, and a fully fledged
account should integrate them both.
Structures and strategies of social cognition
It is against this general background that we also should approach the
question of the structures of ideology, defined as the underlying frameworks
of the socially shared beliefs of group members, as explained in the previous
chapter. Such ideologies are abstract, and hence a more 'structuralise
approach seems more appropriate. Unlike discourse and action, ideologies,
as we understand them, are not locally produced in the sense of shaped by
each specific social context, by a single speaker and utterance (see Chapter
22 for this concept of context). They do not vary from one moment to the
next, and are not strategically adapted to individual recipients. On the
contrary, given their social, group-based functions, they must be relatively
stable and a context free resource for many group members in many
situations. Again, in that sense, ideologies are like grammars, defined as
abstract systems of knowledge (rules) that enable all competent speakers of
a language community to communicate in many different contexts.
On the other hand, ideologies are of course context-sensitive if we use a
broader concept of 'context', including the relevant dimensions of social
56
Cognition
structure, such as groups and institutions, social relations of power, historical development, and so on. Given the earlier definitions, ideologies are
formed and changed as a function of such (broader) social 'contexts',
although such changes are usually rather slow. To avoid further confusion, I
shall not use this broader, commonsense notion of 'contexe, and instead use
the sociological terco 'social structure', or else the terco social 'rnacrocontext' to denote the properties of the social structure that are specifically
relevant for a specific ideology.
That ideologies themselves are relatively stable does not mean that the
expressions and uses of ideologies are not variable, strategic and contextsensitive. On the contrary, the theory will precisely need to spell out how
such expressions of ideologies are adapted by individual social actors and
strategically tailored to the situation at hand. So much so, that they may even
seem to be non-existent in a particular situation. To wit, sexist men will not
continuously make sexist remarks in al situations. Thus, whereas expressions of ideologies in social practices will be variably occasioned and
contextually managed, we assume that ideologies themselves, as well as
other shared social representations, need to be relatively stable.
Such stability is necessary in light of the cognitive and social functions
ideologies have for the many different members of a group in different
situations. Ingroup co-operation, continuity and reliability of action and
judgement and many other properties of successful group membership and
social practices would be impossible without at least a minimum of stability.
In the same way as language users would be unable to speak and understand
their language without a more or less stable grammar, group members would
be unable to accomplish their daily practices and social judgements without
more or less stable social representations such as knowledge, attitudes and
ideologies, of which abstract ideologies are necessarily the most stable
socio-cognitive constructs.
On the other hand, even such more or less stable representations need to
be acquired, changed or abolished by groups and their members, and such
processes of change, though slow, of course need an account of a more
dynarnic nature. That is, all structures, also those of ideologies, eventually
also need an account of their active construction (formation or change) by
group members in social contexts.
Schemata
Whereas structural analysis is a well-known and quite sophisticated procedure in linguistics and discourse analysis, the structural account of cognition
in general, and of social cognition in particular, remains at a relatively
modest level of theoretical sophistication. We have seen that the overall
architecture of the mind is a fairly simplistic construct, with some overall
distinctions between short-term and long-term memory, and between episodic and semantic memory. Beliefs may be represented in (similarly
Structures and strategies
57
simplified) propositions or networks, and belief-clusters may in tum be
organized by various schemata.
This schematic approach is a relatively plain counterpart of structural
analysis in linguistics, and usually lacks the more dynamic dimension
needed to account for the construction, uses or changes of such schemata.
Thus, if we want to explain how people perceive objects, scenes or events,
how they produce or understand sentences and stories, the knowledge they
have to do so is assumed to be organized in such schematic patterns. People
have ideal, abstract or prototypical schemata for the structures of a chair, an
event, a story, people, groups as well as social structures. It has become
standard practice in psychology to specify and distinguish event-schemata,
people-schemata and story-schemata, among others
Such schemata of naive, commonsense knowledge usually consist of a
number of characteristic categories (such as the complication and the
resolution in a story), that may be combined in a specific order and
hierarchy, and allowing for variable terminal elements. Typically, as is the
case in the generative grammar of sentences, such structures are represented
in tree-like (directed) graphs, consisting of a top node, several edges and a
number of lower-level nodes representing subordinate (included) categories.
Note that what is being described here is not real-world objects, but our
socially shared, conventional and cultural knowledge about such objects, that
is, mental structures or representations. It need not be emphasized again that
these structures are merely abstract, theoretical accounts of the organization
of socio-cultural knowledge. Yet, although many alternatives could be
imagined, they should not be arbitrary; they need to account for empirical
phenomena of actual understanding, discourse and action. Some knowledge
structures better account for how people go about perceiving, speaking and
acting than others. For instance, a hierarchical structure may better explain
differences in availability or accessibility of certain top- or high-level
categories than structures that are not organized that way.
However, an account of the organization of the mind that is closer to a
neural model of the brain might provide altemative theoretical accounts that
are based on (neural) nodes or pathways that are in various stages of
readiness or excitation. Theoretically, these may be no more than notational
variants if their descriptive and explanatory power in the account of
information processing, thinking, speaking or understanding is the same.
That is, at lower, more detailed levels of processing, neural models of
representation and processing may be more relevant, whereas at the higher,
more complex level, other representational formats for knowledge, such as
abstract schemata, may be theoretically more useful. 3
The same may be true for actual processing of schemata — at a fairly high
and complex level, people process information linearly, as is the case in the
understanding of words and sentences or the execution of actions. However,
as soon as we want to account for the full complexity of such tasks at all
levels, we must assume that processing needs to be 'massively parallel' as
.a
58
Cognition
the preferred phrase of connectionist theories goes. If we add all the levels
that account for, for example, discourse production and understanding
(phonetic, phonological, lexical, syntactic, semantic, stylistic, interactional,
pragmatic, contextual, etc.), the number of structures being processed in
relation to such beliefs is so high that we must assume that these processes
operate in parallel. Unfortunately, we know as yet very little about the
details of such parallel, neural processing and 'representatioñ when applied
to belief systems.
Scripts
Structural descriptions of social representations may also take a more
dynamic form, especially when they aim to render the structures of events
and actions. Thus, the notion of script has been widely used to account for
the knowledge people have about the stereotypical events of their culture,
such as celebrating a birthday, an initiation ritual, going to the supermarket,
or participating in a university class, among myriad other well-known
events. 4 As the script-metaphor suggests, such knowledge is represented in
terms of a setting, time, location and a sequence of events and actions and
the typical or optional actors that participate in them, like students and
professors in classes, and pilots, flight attendants and passengers in air
travel. Of course, we may imagine other types of structures, as long as they
are able to account adequately for the actual mental and social activities of
people.
It should be emphasized again that such knowledge is general and
abstract. In order to be applicable in the very large number of possible
situations people may be involved in, we must either assume that such
structures themselves are infinitely variable (in a way similar to that in
which the rules of a grammar allow the structural specification of an infinite
number of possible sentences), or that the abstract schemata are being used
by flexible strategies, which may tailor them to each particular situation.
There are also intermediate solutions, where schemata or scripts are assumed
to be built up of smaller structural units (in the way that 'paying' is a
sequence of basic actions that may be found in most economic interactions,
like buying a product, or paying for a ticket in the movies) that may be
combined and hence vary in a more flexible way. 5 But even then, actual
variation is practically infinite, given the (theoretically) infinite ways of
accomplishing these component basic actions. So, any account, whether a
more structural or a more strategic one, has or needs to be complemented
with flexible rules or strategies that adapt structural categories or units to
their variable uses by different people in different situations. This is as true
for the production and comprehension of sentences, as it is true for everyday
conversations, complex institutional dialogues, or for more or less complex
social acts such as going to the movies, managing a firm or governing a
country.
Structures and strategies
59
The main point in all these cases is (a) that we need to assume socioculturally shared and mentally represented knowledge, (b) that such knowledge needs to be organized so that it can be effectively acquired, accessed
and changed, and (c) that such knowledge needs (internal or externa»
strategic means for its variable and effective uses by individual users in
concrete situations. Later we shall see that we need to add a number of social
properties of knowledge — it is not acquired, used and changed in abstract
situations, but in social situations by social actors, as well as in institutions,
organizations and whole cultures.
Organizing evaluations
With all their theoretical limitations (most schema theories are not exactly
examples of formal explicitness and conceptual sophistication), these various approachés to the account of the structures and strategic uses of
knowledge have been relatively successful. It is not surprising, therefore,
that similar schema-theoretic roads have been followed in social psychology. 6 Thus, if people have schemata or scripts for stores, stores, stories and
storytelling they probably also have them for people, groups, intergroup
relations, domination, organizations, governments and democracy. The same
is true for the myriad of communicative events that describe or constitute
such social objects, such as conversations, negotiations, parliamentary
debates, impression management as well as corporate management.
The theoretical task then consists in spelling out these various structures
as well as the strategies of their usage. This is easier said than done. One
question is whether it is likely that all or at least many of these mental
representations have the same or similar categories or whether their overall
structures are at least the same or similar, if only because of obvious reasons
of cognitive economy. Intuitively, we may assume that there are considerable differences: our beliefs about chairs, chairpersons and chairing a
meeting probably do not have the same internal organization. Yet, chairs
may have structures that are at least comparable to many other objects,
chairpersons are not very different from other people or roles, and chairing a
meeting is not essentially different from many other forms of interaction. So,
we may have object-schemata, person-schemata, role-schemata, and
interaction-schemata, and similar schemata (or scripts, or scenarios, etc.)
maybe developed for groups, relations of domination, organizations, civil
wars, democracy or, indeed, ideologies.
However, there are some complications. What has been said, aboye,
especially applies to the organization of knowledge, but does it also apply to
the organization of opinions, attitudes and judgements? We may postulate
person-schemata and group-schemata, and maybe scripts for parliamentary
sessions and civil wars, but how do we organize the opinions and attitudes
we have about such social objects or events?
Despite a number of modest attempts, 7 few detailed representation formats have been provided for evaluative structures. In fact, we do not even
60
Cognition
know whether such evaluations should be represented separately from our
knowledge about the objects of judgement. If people have a group-schema
about, say, Turks, does this mean that such a schema should also feature the
opinions and the prejudices people may have about Turks?
For instance, a simple network could have 'Turks' as a node, and this
node would be related to nodes that specify our knowledge about Turkey as
a country, about Turkish as a language, about Turkish society and culture
and so on, but that central node would also be related to nodes representing
our evaluation of Turks as a people (or about the Turkish language, culture,
religion, etc.). If many or most of the important (or central) nodes of the
Turk-schema or Turk-network were negative, then this would represent a
prejudice. Such a simple, integrated approach, where factual and evaluative
beliefs are represented in one group-schema, meets a number of criteria for
cognitive organization, namely, those of simplicity and economy. The
question is whether it works — does such a schema account for prejudiced
discourse and interaction, and does it explain discrimination, among many
other forms of biased perception and interaction?
Attitude structures
Although at present we doñ t have a clear answer to such questions, we may
however take a different theoretical approach and assume that in the same
way as factual and evaluative beliefs can be distinguished, we may also
distinguish between factual belief structures, on the one hand, and evaluative
belief structures on the other hand. At the moment, this is merely an
analytical distinction: it may very well be that in the mirad (in the brain)
these form one network. But, following the common sense of social
members, we may provisionally distinguish between cultural knowledge, on
the one hand, and group knowledge and group attitudes on the other hand.
One argument for this separation, apart from differences in social practices and discourse, is that knowledge is socio-culturally based on different
methods of assessment and verification, namely, truth criteria such as
observation, reliable sources, argurnentation, proof or experimentation.
Opinions are constructed and combined along very different methods of
assessment, and following different criteria, such as values, group goals and
interests, and social group relations. To establish where Turks come from in
the world, what language they speak, or what religion they have, among
other things, requires 'informatioñ from newspapers, textbooks, atlases,
everyday conversation and observation, as well as inferences from other
knowledge, for instance about languages, religions, Islam, politics or the
Mediterranean. ,When expressing such knowledge, as such, language users
presuppose that others have similar beliefs (truthful or not) and that the
methods to establish the truth of such beliefs or to setde disputes are socioculturally shared.
However, prejudices about Turks are developed and used, and probably
organized in quite a different way. First of all, as empirical evidence shows,
Structures and strategies
61
people may have negative attitudes about Turks even without having any
knowledge about them. Indeed, experiments and everyday experiences show
that some people even express prejudices about non-existent peoples! And
although most people who have prejudices about other groups usually have
at least some knowledge about such groups, knowledge often prevents
stereotyping and prejudices. The development of prejudice precisely avoids
the methods and reliability criteria of knowledge, such as repeated observation, inference, proof, reliable sources, and relations with other knowledge;
hence, obviously, their role as forros of pre-judgement. Generalizations are
made from one or two observations, fallacies made in argumentation,
unreliable sources are used, if at all, and so on. 8
Even more importantly, apart from such'falliblé information processing
and judgement (which characterizes much thinking in general), what counts
in the construction of prejudices are the goals, the interests and the values of
the group. That is, if the group is Christian, and if Islam is defined as
different, opposed to, or even as a threat to Christianity, and hence to Us,
then Turks, fike most other Muslims, may be negatively represented on the
relevant category of religion. The same may happen for appearance, origin,
employment, language, habas or perceived personal traits. In other words,
besides the relevant knowledge categories for groups, group members may
bring to bear a number of categories that are (for them) essential in the
evaluation of other groups. One of these categories may be appearance, so
that, for white people, anybody who is not white (and having other features
of'Europeañ appearance) may be categorized as essentially different,
deviant or dangerous on that dimension. Whether such basic categories have
historical or even biological foundations is irrelevant. What counts is their
socio-cultural construction and reproduction. People may learn and unlearn
that differences of appearance are crucial in categorizing and especially in
judging others.
The point then is that in general the structures of evaluative social
representations such as attitudes (and as we shall see, of ideologies) are
probably organized in a way that reflects or facilitates their social (groupbased) functions, their social construction, and their social uses in everyday
social practices. ff skin colour is relevant to categorize and judge negatively
other groups in order to be able to discriminate against them or oppress
them, then such a real (or indeed imaginary) characteristic may become a
category of the evaluative schema that defines (ethnic) attitudes in general
and prejudices in particular.
Traditional approaches in the social psychology of attitudes follow some
of these arguments in their assumption that attitudes always consist of three
components: a cognitive, an evaluative and an emotional one. 9 Obviously, a
three-component assumption does not tel us much about detailed structure
or organization, only about the nature of the beliefs involved in attitudes. I
have further argued that, whatever the 'real' organization of beliefs in the
brain—mind, I prefer to keep factual beliefs apart from evaluative beliefs, and
62
Cognition
hence distinguish between knowledge and attitudes. As defined, the latter are
only evaluative.
Finally, since emotions (when not confused with evaluations) are strictly
personal and contextual, they cannot be part of socially shared, abstract
group attitudes. They may, however, become triggered and mingled with the
actual uses of attitudes in concrete situations by individual members. I may
now be angry (or desperate) about a political decision, an emotion that may
be triggered by activating or constructing a negative opinion in the current
context. But a group cannot be continuously 'angry', in the strict sense of
being angrily aroused. Socially shared, continuous 'affece, such as hate or
anger, is not, in my view, an emotion, but a forro of strong evaluation (which
may of course be expressed in the language of emotions). It is highly
unlikely that there are groups all of whose members are constantly emotionally aroused about some issue, but as is the case for ethnic prejudices, they
may well share and maintain a negative evaluation about other groups.
Following a more fruitful way of cognitive inquiry into the more detailed
organization of evaluative belief clusters, I assume that group members
develop schemata or other abstract structures for the organization of social
judgement. Such attitude-schemata for groups, thus, will feature those
general categories that have developed as a function of the goals, interests as
well as the social and cultural contexts of group perception and social
practices. In some socio-historical situations this may be skin colour (as with
prejudices of whites against blacks), religion (as in anti-semitism), gender
(as in sexism), political ideology (as in anti-communism), and so on. Thus,
whatever is relevant for evaluation, and the practices legitimated in tercos of
a negative (or positive) evaluation, may thus be selected as a category of the
group-attitude-schema.
These schemata may be different for different types of group relationships, narnely, those based on origin, ethnicity, gender, age, class, profession, and so on, but the same principies will be at work in the
construction of attitudes. Note again that although it is plausible that both
knowledge and attitudes usually operate in the conduct of discourse and
other social practices, attitudes are distinct from knowledge, and so are their
internal structures. Categories in attitudes may be used that have no basis in
knowledge at all, but that are simply useful for negative judgement. The
same is true for the order or relevance of such categories in the scherna.
Thus, in ethnocentric and racist attitudes, the appearance of other group
members (even when 'objectively' barely different from that of our own
group) may take the highest position in the category, and the same may be
true for language, religion, socio-economic status, occupation, habits or
attributed personal 'character' (e.g. being lazy or criminal)
Interestingly, and as we shall see in more detail later, the selection and
ordering of categories of judgement are obviousiy not arbitrary, but a
function of the social position, goals, resources, activities and other interests
of the group that shares such an attitude. For the unemployed 'They take
away our jobs' may become a prominent judgement of a prejudiced attitude,
Structures and strategies
63
so that the socio-economic position of the other group becomes crucial. This
process not only plays a role in relations of domination, but also of
resistance. Thus, for linguistic minorities, the language of the linguistically
dominant group will be a major category of judgement.
What seems rather straightforward for the organization of social opinions
about other people and groups — the construction of evaluative group
schemata consisting of variable hierarchies of social categorizations — is less
obvious for attitudes about social issues and problems, such as abortion,
nuclear energy or pollution. Although here also groups of people are
involved about whom we may develop opinions, such attitudes rather focus
on right or wrong social practices, or even about properties of objects or
nature. Semantically such 'problems' may be construed (by different groups)
as implying some kind of norm violation, if not as a threat, but such
semantic contents are not readily reduced to abstract, general categories that
allow the description of large classes of attitudes. And yet, given the
typically organized nature of the mind, it is highly unlikely that such
attitudes merely consist of lists of propositions representing opinions about
what people like or dislike.
My theoretical approach tries to go beyond the traditional approaches to
the structures of opinions in social psychology, such as consistency and
balance theories. What we find here is an account of the mutual relations
between (sets of) propositions and the dynamics of their acceptance or
rejection by individuals. Thus, adopting mutually inconsistent opinions may
create 'cognitive dissonance', which people try to resolve by strategically
adapting their opinions. Similarly, we may find further analysis of opinion
propositions in evaluative 'molecules' whose development and change may
mutually influence each other. If, for instance, I like John but disapprove of
nuclear energy, than what happens when I also know that my friend John
does approve of nuclear energy? Would this make John less likable and/or
nuclear energy less detestable, or do I apply other useful strategies to
combine the inconsistent 'valences' of my opinions? 1.
These traditional questions about the acquisition, organization and change
of opinions and attitudes remain relevant today. However, they address
somewhat different dimensions from those I am interested in. First, they do
not distinguish between personal and social opinions, nor indeed between
opinions and attitudes. Secondly, they focus on the individual 'management'
of opinions in specific contexts and situations, rather than on general,
complex and socially shared attitudes. Thirdly, they do not answer the
question about the oyeran organization of such attitudes, and the relations of
such an organization with the social dimensions of the groups that entertain
them. However, such questions are still relevant as soon as we need to
examine the ways concrete opinions are produced by individuals in specific
contexts, possibly as a result of mutually 'inconsistent' attitudes. These
strategies of opinion management and the representation of opinions in
mental models (see Chapter 7) need to be dealt with separately.
64
Cognition
From this discussion, we may provisionally conclude that evaluative
social representations, such as attitudes, have their own , that is, their
own, socially based schematic organization and categories, which are a
function of the symbolic or material interests of the group. I shall later
examine in more detall what these 'interests' are.
The argument I have been pursuing in this chapter suggests that if all
social representations have their specific structural categories and organizational principies, this should also be so for the very foundation of such social
representations, that is, for ideologies. This hypothesis will be explored in
the next chapter.
5
Structures of Ideologies
Searching for a format
Given the assumption that social representations such as knowledge and
attitudes of groups are organized by a non-trivial structure, it is plausible
also that ideologies are not merely a list of basic beliefs. The acquisition, the
changes and the uses of ideologies in social practices suggest that we should
try to find schemata or other structural pattems that are typical for ideological systems. Since we have no a priori or theoretically obvious format for
such structures, we have to build such schemata from scratch and find
evidence that suggests how ideologies may be organized.'
One heuristic option is to assume that the structures of ideologies are
similar to those of other social representations. For instance, if scripts
organize our knowledge about stereotypical events, do ideologies also have
such a script-like nature This assumption may be rejected without much
hesitation: whatever we know about ideologies, they do not in any way
reflect the stereotypical structures of events. First, ideologies are much more
general and abstract, and do not merely apply to specific (types) of cultural
events, such as shopping or going to the movies. Second, ideologies not only
apply to events, but also to situations, processes, groups, group relations and
other facts. Indeed, given the fundamental nature of ideologies and their
assumed role in the management of social representations of groups and
group relations, they should somehow reflect how groups and their members
view a specific issue or domain of society. Third, ideologies do not merely
control knowledge but also opinions about events, and such opinions do not
represent event structures. Scripts, therefore, do not constitute a likely
candidate for the kind of organization we would expect ideologies to have.
Since attitudes are clusters of socially shared, evaluative beliefs, it is
therefore more plausible to examine whether ideologies have the structural
features of attitudes. Such an assumption would probably also make it easier
to link ideologies with attitudes, for instance when we assume that ideologies organize attitudes, or that they assign some forro of coherence to the
clusters of attitudes that are govemed by the same ideology.
Especially since we as yet have no definite idea about what attitudes look
like in general, our question about the similarity of attitude structures and
ideological structures might well be rnoot. So, let' s take a few examples of
attitudes and see whether their possible structures suggest a more general
format that also may be relevant for ideologies. For instance, there is good
? 2
66
Cognition
evidence that at least some groups of people have attitudes about immigration, abortion and nuclear energy. Thus, a (prejudiced, nationalist or racist)
attitude about immigration may feature the following evaluative beliefs,
among many others:
1 Too many people come to our country.
2 Our country already has too many people.
3 Immigrants only come here to live off welfare.
4 Most immigrants are economic refugees.
5 Immigrants require scarce housing and jobs.
6 Immigrants face growing resentment in the inner cities.
7 The government must send back illegal immigrants.
8 Immigration has to be restricted to 'real' refugees only.
These evaluative beliefs, which are routinely expressed in both elite and
popular discourse about immigration, together define the (negative) attitude
about immigration? However, as presented, it merely seems to have the
structure of a list of beliefs. If there is structure here, it is at rnost an
argumentative one: opinions 1 to 6 may be interpreted as arguments that
support the normative political conclusions 7 or 8.
At another level of abstraction, attitudes may be structured by the basic
categories ofproblem and solution, where the problem category is recursive.
Immigration is conceived of as a set of problems that result from immigration: overpopulation, lack of housing and work, growing resentment, and so
on. The solution category, what must be done to solve the problem, in this
case coincides with the main normative conclusion of the other opinions.
This specific example does not imply, incidentally, that all ideologies and
attitudes have a problem/solution structure. However, many ideologies,
especially of dominated and dissident groups, organize around basic beliefs
about what is wrong, and about what should be done about it.
If we were to disregard the general nature of the beliefs (this attitude
exists in most European countries as well as in North America), it could
even be organized as a story, with an orientation such as 'Our country did
not have many problems and not many immigrants'; complication: 'Suddenly many immigrants carne to the country, and caused a lot of social and
economic problems; resolution: 'Restrict the number of immigrants.' 5
Finally, some further structure may be assigned to this attitude by
applying a group-schema to it, in which immigrants are characterized by, for
example, the following categories and their (here highly simplified) belief
contents typical for a prejudiced attitude: 6
• Origin: Third World;
• Appearance: mostly people of colour (unlike Us);
• Socio-economic characteristics: they are poor and
want to become
rich;
• Cultural characteristics: they speak other languages, are often Muslims,
and have strange habits;
Structures of ideologies
67
• Personal characteristics: they are illegals/criminals, cannot be trusted,
doñ t want to work hard, etc.
We see that attitudes may be organized in different ways: in tercos of an
implicit argument, in terms of problem/solution categories or the related
categories of stories, and finally in tercos of a group schema.
Further analysis, however, suggests that these structures can only be very
tentative. First, the most articulate structure, namely, the group schema,
defines an attitude about immigrants rather than about immigration, although
these attitudes are of course closely related. Second, narrative and argumentative structures characterize the discourse in which these beliefs may be
used, but not the beliefs themselves. 7
The problem/solution category seems more promising, since it is very
abstract and general, and reflects the fact that attitudes are usually developed
for social issues or problems, as seen by a specific group. For the groups
who share them, the same is true for evaluative beliefs about nuclear energy
or abortion. Yet, this structure is so general that it has litde organizational significance, since it does not say more than that a social issue is a problem for the
members of the group, and that these members also have a solution for it.
Do ideologies have a problem/solution structure? Many ideologies indeed
seem to have something like that. Thus, whereas racism typically defines
immigrants, foreigners, minorities or others as the reason for most social and
economic problems, and withholding 'our' scarce resources (residence,
citizenship, housing, employment, equal rights, etc.) as the solution, similar
simple analyses may be made for anti-racism (problern: racism; solution:
equality, diversity, etc.), feminism (problem: male chauvinism; solution:
equal rights, etc.), and environmentalistn (problem: pollution; solution: stop
polluting). Other ideologies, such as liberalism, do not seem to have such a
clear problem/solution structure, although originally it may have had such an
organization as an opposition ideology against feudalism.
In sum, where attitudes seem to represent a problem or a social conflict,
they may well have at least some structural features that we also find in
ideologies. This is of course hardly surprising since ideologies are most
likely to represent (real or imaginary) problems and conflicts of interests of
— or between — social groups. As is obvious from the example of immigration, there is therefore also a strong polarization between Us and Them, as
representatives of the groups involved in such a conflict. Similar observations hold for the attitudes about nuclear energy and abortion.
Very tentatively, these examples provide us some suggestions for at least
some ideas about the format of ideologies: problem/solution, conflict and
group polarization. Let us analyse these potential categories of ideological
structure in more detall.
Group conflict
Although ideologies may have some features that we also encounter in more
specific attitudes, we need to explore a bit further to come up with a format
68
Cognition
that is general enough to fit afl ideologies, and specific enough to be nontrivial and functionally useful in the cognitive management of ideologies as
well as in their acquisition and applications.
Instead of starting with the organization of social representations in
general, we may also inquire whether the structure of ideologies is a function
of their role in society. We have already seen that often social conflicts
between groups with different interests are involved. We also know from
most traditional approaches that ideologies are typically used as foundations
for domination and resistance; that is, they represent social struggle. Moreover, ideologies are also intuitively functioning as self-serving principies
involved in the explanation of the world in general (as in religious
ideologies) and of the social and economic worlds in particular (such as
conservatism or capitalism) Finally, ideologies have a normative dimension,
and summarize what group members should do or not do: for example resist
oppression, stop pollution, or prevent abortion.
If we assume that many if not most ideologies are a socio-cognitive
representation of the basic evaluative and self-serving beliefs of group
members about social struggle and group conflicts, it may be most fruitful to
study this fundamental feature in more detall in order to find out the most
effective format that might organize such beliefs. Crucial for such a
representation is how group members see themselves and how they see
Others.
Thus, typical for a racist ideology is that we are representing Us as
superior, and Them as inferior, and that as a consequence we (should) have
preferential access to society's scarce resources (for an empirical case study
of such a racist ideology, see Chapter 28). This is even the case when racist
groups claim Us and Them to be equal but different, and hence advocate
separation of the 'luces', because also in that case no equal access to scarce
social resources is usuálly permitted. A similar basic representational format
may be postulated for male chauvinists and their opinions about gender
relations. Feminist ideologies are not merely the mirror-image of sexist
ideologies, but represent Them (men) as oppressing Us, and themselves as
engaged in resistance against gender inequality. Religious ideologies represent Us as (good) believers and Them as (bad) non-believers (infidels,
heathens, etc.). And finally, environmental ideologies represent Them as
polluters, and Us as those who resist pollution and defend nature and the
rights of animals, for instance. More generally, conservatives see themselves
as defending traditional social relationships and moral values against Them
(progressives, etc.) who want to change these in favour of social equality.
Recall that these highly simplified ideological representations are not, as
such, true or false, although each group will of course tend to believe its own
ideological beliefs to be true or justified. Thus, we may agree that prejudices
based on racist or sexist ideologies are wrong or otherwise misguided, and
hence defined in negative terms, but this evaluation of course only holds on
the basis of an anti-sexist or anti-racist ideology.
rl
Structures of ideologies
69
The very general polarization schema defined by the opposition between
Us and Them suggests that groups and group conflicts are involved, and that
groups build an ideological image of themselves and others, in such a way
that (generally) We are represented positively, and They come out negatively. Positive self-presentation and negative other-presentation seems to
be a fundamental property of ideologies. Associated with such polarized
representations about Us and Them, are representations of social arrangements, that is, the kinds of things we find better (equality, a clean
environment) or those which we believe others stand for (inequality, a
polluted environment, a free market). At this very abstract level these social
arrangements are specifications of more general values.
Thus, if'freedom' is a general, socio-cultural value, then 'freedom of the
market' is one of the things a capitalist ideology will represent as something
We stand for; feminista will translate this general value in terms of the
freedom of women (freedom from oppression and inequality, freedom of
choice, and so on); and environmentalists will interpret the value as freedom
from pollution, and so on. We shall focus on the nature of values later, but
they obviously play a fundamental role in ideologies. This is not surprising
when ideologies are taken to be the basis of group beliefs.
In sum, ideologies are representations of who we are, what we stand for,
what our values are, and what our relationships are with other groups, in
particular our enemies or opponents, that is, those who oppose what we
stand for, threaten our interests and prevent us from equal access to social
resources and human rights (residence, citizenship, employment, housing,
status and respect, and so on). In other words, an ideology is a self-serving
schema for the representation of Us and Them as social groups. This means
that ideologies probably have the format of a group schema, or at least the
format of a group schema that reflects Our fundamental social, economic,
political or cultural interests.
Such an assumption is plausible when we think of the various social
functions of ideologies, to which we shall return in more detall later. Thus,
ideologies may be used to legitimate or obscure power abuse, or conversely
they may be used to resist or denounce domination and inequality. Ideologies thus are needed to organize our social practices in such a way that
they serve our best interests, and prevent others from hurting such interests.
These various more or less intuitive conceptions of the nature and
functions of ideology, and the assumption that ideologies may be represented as group schemata, suggest the following categories for a tentative
format of the structure of ideologies:
Who are we? Where are wé from? What do we look like?
Who belongs to us? Who can become a member of our group?
• Activities: What do we do? What is expected of us? Why are we here?
• Goals: Why do we do this? What do we want to realize?
• Membership:
1
70
Cognition
What are our main values? How do we evaluate ourselves
and others? What should (not) be done?
• Position and group-relations: What is our social position? Who are our
enemies, our opponents? Who are like us, and who are different?
• Resources: What are the essential social resources that our group has or
needs to have?
• Values/norms:
These categories and the basic questions they stand for seem to be the
fundamental co-ordinates of social groups, and the conditions of their
existence and reproduction. Together they define both the identity as well as
the interests of the group. Thus, if ideologies are primarily representations of
the basic properties of groups, then this schema should be a serious
candidata for the organization of ideological beliefs.
This schema seems fairly generally applicable to all ideological groups,
whether based on more or less inherent characteristics (gender, ethnicity,
age, etc.), on what we do (as for professional ideologies), our goals (as for
ideologies of action groups), norms and values (as for conservatives versus
progressives; religious and non-religious people), our relations with others
(superiors versus subordinates), and the typical resources we do or do not
have (rich versus poor; employed versus unemployed; homeless versus those
who have a home). That is, each category may be needed to define all
groups, but groups may also be identified specifically by one particular
category.
This may also explain why there are differences between membership,
activity, goal, etc. ideologies. Thus, feminism is typically a goal ideology,
that is, defined by the hierarchically most important belief of the ideology,
namely, to arrive at full equality for women and men. Similarly, the
ideology of black nationalism is a membership ideology when it is limited to
questions of appearance and 'racial pride' (as old slogan about 'black is
beautiful' and 'négritude' imply), and a position or resistance ideology
when it focuses on self-determination and black empowerment. Capitalism
on the other hand would rather be a resource ideology, aiming to ensure
freedom of enterprise and freedom of the market. In other words, the
categorial structure of ideologies also allows a typology of ideologies, as
well as the possibility of changing hierarchies in the representation of
ideological beliefs.
Each category of this ideological format functions as the organizing
pattem of a number of basic evaluative beliefs. Note though that these
beliefs are by definition ideological. Thus, journalists in their professional
(activity) ideology, may represent themselves essentially as gathering and
bringing the news, for instance. They do this, they would say, in order to
inform the public and more generally to serve as a watchdog of society.
Obviously, these are ideological goals, because we know that many journalists hardly do this. That is, such a goal is at most a benchmark or a property
of an ideal type: how journalists would like to be. The same is true for their
(professional) values, such as truth, reliability, fairness, and so on. The
1
Structures of ideologies
71
specific resource of journalists, access to which must be guaranteed as a
condition for the existence or the activities of the group, would be information or the freedom of the press (as it is freedom of the market for managers,
and freedom of research for scholars, and freedom from discrimination for
feminista and anti-racists).
It should be emphasized that this abstract categorial schema is merely a
theoretical construct that may .be used to organize and explain the basic
evaluative beliefs of group members. This schema, as such, does not tell us
yet how ideologies are acquired, used or changed, how they manifest
thernselves in social practices, and how they reproduce thernselves in
society. It is also a social representation. This means that it characterizes
groups, at a macro-level. Individuals members on some dimension may not
identify with the group, and hence not share the ideology of the group.
Socially this usually implies that they are considered as dissidents, traitors,
deviants, or otherwise as group members who no longer 'belong' to the
group, and may hence be excluded, marginalized or otherwise punished. I
shall return to these and other social conditions and consequences of
ideological group memberships later.
Note that, at the moment, the schema primarily serves as an organizing
frarnework for ideological beliefs. That is, its function here is cognitive. Yet,
as suggested, each of its categories is rooted also in social structure, that is in
group membership criteria, social activities and goals, group relationships,
social values and social resources. This will later allow us to define
ideologies precisely as the socio-cognitive interface between the (mental)
social representations shared by the group, and the social identity, activities,
organization, and so on, of the group and its members.
Later I also need to analyse how this abstract schema, designed as an
organizational pattern for ideological beliefs, can be empirically founded.
That is, we should see it not just as a theoretical construct, but as a schema
that actually does play a role in the acquisition, changes and uses of
ideologies. One of the ways to assess the empirical nature of the schema is
to make a systematic study of social practices, and especially of discourses
that express ideological beliefs. If these expressed beliefs and their inferences appear to be organized according to the ideological schema, then we
have some evidence that the schema is indeed a socio-cognitive device used
by social groups and their members to organize their basic beliefs.
There is an interesting iniplication of choosing a group schema as a
format for the structure of ideologies, namely, the obvious relation it has
with group identity. If ideologies monitor the way people as group members
interpret and act in their social world, they also function as the basis of their
social identity. Structurally this would suggest that the first category
(membership) is not the only one that defines identity, although it seems to
organize beliefs about what we 'essentially' are (white, black, men, women,
poor or rich). However, it is obvious that the whole schema, all categories
together, defines group identity: what people do, their goals, their values,
their relations to other groups, and their resources for survival or social
72
Cognition
existence also are part of their identity. The first category in that case defines
only a fragment of group membership: a number of more or less inherent or
relatively permanent properties (such as origin, appearance, gender, religion,
language or other cultural specifics) that define primary group membership
criteria, as well as conditions of inclusion and exclusion (for details about
social identity, see Chapter 12).
Also, it should be stressed from the outset, as will be developed in more
detall later, that social actors are obviously members of many social groups,
and that therefore they have multiple, sometimes conflicting identities and
hence share a mixture of ideologies. Discourses and social practices in
concrete contexts will show such complex combinations, conflicts and
sometimes inconsistencies. The same is true, cognitively, for people' s
attitudes, models and opinions, which may be monitored by different
ideologies, of which the unique combination may be personal or limited to
sub-groups (such as the sub-group of US middle-class black women journalists). Obviously, empirical research needs to take such complex interactions
into account in order to be able to describe ideological social practices and
discourse (for an illustration of how several ideologies interact, see Chapter
28). 8
Contents
The same is true for the contents of the respective categories of the schema.
What we now have is an abstract framework. Ideologies, however, are
content-specific, and further empirical work is necessary to spell out the
actual group beliefs that are organized by these categories. This will also
allow us to link the ideologies with the more specific attitudes that are in
turn controlled by these ideologies. I shall therefore be brief about the
contents of ideologies.
At an elementary level of analysis, ideologies consist of clusters of basic
social beliefs organized by the schematic categories proposed aboye.
Although these beliefs may in principle be about anything that relates to the
social experiences and practices of social groups and their members, they
will mostly be about conflicts of interests between groups, typically so in
relations of competition, domination and resistance. That is, ideologies
usually organize attitudes which in turra control those social practices of the
group and its members that are somehow relevant to the interests or identity
of groups, and are related to membership criteria (inclusion and exclusion),
activities, goals, values, relations to other groups, and resources. Since these
beliefs are often evaluative, they presuppose socio-cultural values, such as
truth, co-operation, equality, freedom and autonomy, among many others
(see Chapter 6). Thus managers may hold the ideological belief that they
want to be free from state intervention, and feminista that they want to have
equal rights as men, among many other ideological beliefs.
In sum, the contents of group ideologies pertain to what, for each group,
is the preferred social and moral érder, whether or not such an order is seen
Structures of ideologies
73
as just or unjust. Yet, although it may seem as if some groups may develop
ideologies which 'cynically' acknowledge that they are not 'just' for other
groups, the fundamental social role of positive self-images for most groups
will usually entail that groups develop an ideology which they see as
ethically good or defensible. Thus, whereas (at least some) proponents of a
neo-liberal ideology may recognize that liberalization and other market
policies may make the rich richer and the poor poorer, it is likely that the
underlying ideology maintains that freedom of the market will eventually
benefit all. In that respect, we may generally assume that group ideologies
cannot be 'cynical', but always imply positive self-presentation. One possible exception that needs to be further explored may be ideologies of some
dominated groups, a 'false consciousness' that may result from manipulative
hegemony, in which the own group is represented as negative in relation to
dominant groups, as would be the case for forms of interiorized racism.
Ideologies develop as a functional consequence of the conflicts of interest
that emerge from goals, preferences or rights that are seen as mutually
incompatible. Groups may want to claim, defend, legitimate, explain, or
otherwise manage such interests against other groups in society, while at the
same time rallying their own members behind such claims so as to make
sure that attitudes of individual members, and social practices based on
them, co-ordinate and facilitate the realization of ideological goals. In the
remaining chapters, I shall further examine such ideological 'contents', and
especially their social conditions, consequences and functions in the management of social interpretations, practicés and discourse.
Values
Introduction
Values play a central role in the construction of ideologies. Together with
ideologies they are the benchmark of social and cultural evaluation. Like
knowledge and attitudes, they are located in the memory domain of social
beliefs. That is, we do not take values as social or sociological abstractions,
but as shared mental objects of social cognition.
Unlike group beliefs, values have a broader, cultural base. Together with
culturally shared knowledge, they are part of the cultural common ground.
Whatever the ideological differences between groups, few people in the
same culture have very different value systems — truth, equality, happiness,
and so on, seem to be generally, if not universally shared as criteria of action
and at least as ideal goals to strive for. Of course, there are cultural
differences. Some values may not even exist in another culture, or may have
different implications in another culture. Also, the hierarchy of the imponance or relevance of values may be different from culture to culture.
Whereas in one culture honesty may be fundamental, another culture may
emphasize modesty. For these reasons, cultural clashes and conflicts of
values, also in communication, are notorious, as is especially clear in
variations in politeness, deference or directness of text and talk, among
many other differences. 1
Values are shared and known, and applied by social members in a large
variety of practices and contexts. Obviously, they form the basis of ah
processes of evaluation, and hence for opinions, attitudes and ideologies.
Thus, if ideologies are the basis of group beliefs, and if values are in turra
broader and more fundamental, values must be the basis of the evaluative
systems of a culture as a whole. Indeed, values are the pillars of the moral
order of societies.
This fundamental socio-cultural status of values also precludes their
reduction to individuals. These may share, adopt or reject the values of their
group, but we would not say that personal goals or ideals are values.
Value systems
Despite the frequent use of the notion of value in the social sciences and
politics, they are fairly elusive. Usually, and unlike beliefs, they are
described in isolated terms, such as truth, intelligence or beauty, or in terms
of concepts for which English does not have a single word, such as
Values
75
'enjoying a good life'. If they are basic building blocks of the evaluations
that are involved in social opinions, as attributes that are predicated of any
socially relevant object (people, events, actions, situations, etc.), then they
might well be atomic concepts. Indeed, truth or beauty hardly seem
decomposable in more elementary concepts, unless these would be good and
bad, so that 'beauty' would be 'good appearancé , for instance, whereas
honesty would be one type of 'good character'.
This attempt at analysis also suggests that values seem to be organized by
the fundamental dimensions of everyday life experiences and observation as
well as for social action and organization. Thus, we have actitudes that
describe positive properties of the mirad (intelligence, smartness, erudition
and wisdom), whereas others characterize what we value most about bodies:
health, beauty, and so on. Similarly we have a series of values for
judgements about personal 'character', such as honesty, integrity, modesty,
kindness, openness, patience and so on.
The same is true for actions, which also need to be evaluated routinely,
and hence require a complex set of values, such as resolution, decisiveness,
speed or efficiency. Interaction requires evaluation by means of a series of
social values, such as politeness, tolerance, co-operation, helpfulness or
altruism, among many others. As elsewhere, the opposed concepts by
definition denote negative evaluations of people, that is, what people
generaily would not want to be or do, or be accused of: impoliteness,
intolerance and egocentrism. Many of the interaction values mentioned here
of course also apply to discourse, as is obviously true for politeness and cooperation.
What is true for action and interaction, also applies to more complex
social structures, social relations, organizations and whole societies. This
means that democracy, freedom, equality, independence or autonomy are
such fundamental societal values. Given the nature of ideologies as basic
systems of group beliefs, we may assume that these typical societal values
play a special role in them, as is indeed the case — virtually all major social
and political ideologies will emphasize one or more of these societal
values.
In sum, if we draw an intuitive picture of the personal and social world,
each fundamental dimension (mirad, body, character, action, interaction,
society) of observation and evaluation has its own special values. Some of
these values may be very general and apply across these dimensions, as is
true for good and bad, ugly and beautiful.
Finally, the interpersonal and social scope of cultural values probably
does not exhaust the system — we also have values to qualify nature or
animals, either in very general tercos of beauty, but, as is obvious in
environmental ideologies, also in terms of cleanliness, being unspoiled, and
so on. The same is true for all objects of our senses, so that for our taste
alone we have a long series of values: sweetness, delicateness or smoothness, obviously also culturally variable.
76
Cognition
These examples also show that many values are historical: They were
once Invented' as being positive properties of mirad, action or society 'we'
would have to strive for. This 'terminal' aspect of ideologies also suggests
that they are motivational and goal-oriented; that is, they qualify'ideal' endstates or results of human endeavours.
Values and ideologies
Theoretically, then, values monitor the evaluative dimensions of ideologies
and actitudes. That is, basic social opinions are constituted from values when
applied to specific domains and issues in society. Thus, if journalists value
truth and reliability of reporting, then this is an ideological specification of
the cultural value of truth and reliability. The same is true for the selection of
the value of equality by feminists, minority groups or anti-racists in the
construction of their egalitarian ideologies.
Obviously the selection and construction process by which values are
incorporated in ideologies is again self-serving. It should fit the various
interests of the group, such as their membership, activities, goals, relations
to other groups and resources. In other words, general cultural values may be
'appropriated' by a group, as is typically the case with freedom in neoliberal and conservative ideologies. This is also why the values category
itself was added to the ideology schema proposed in the previous chapter:
The values selected as primordial for each group constitute the selected
benchmark for their identity and self-evaluation, the evaluation of their
activities and goals, and especially their evaluations of other groups,
underlying goals and judgements of interaction.
Also negatively, values may be used for self-enhancement, as when white
racists feel superior to non-whites. This superiority feeling is a summary of
a biased comparison process in which We are seen as more intelligent, more
efficient, harder working or more democratic than They are. That is, for all
values that are especially relevant to us, we self-evaluate Us as better. At
most we may grant them superiority on values that are less relevant for us,
such as musicality, being good in sports or hospitality.
As we have seen for the example of freedom before, different or even
opposed social groups may select the same value, but invest it with very
different ideological content. Managers ideologically 'incorporate' (pun
intended) the value of freedom as 'freedom of the market' or 'freedom from
state interventioñ , as a self-serving ideological goal that guarantees their
power and interests. Journalists similarly want to secure their power,
interests and resources by emphasizing the freedom of the press, or the
freedom of information, obviously primarily for themselves. On the other
hand, liberation movements, feminism, and other dominated groups focus on
freedom as a guarantee for equality, independence, autonomy and access to
scarce social resources, and generally as, 'freedom from oppressioñ .
We see that the positive values that define the moral order of a society or
culture are used by all groups not only as a criterion of evaluation, but also
Values
77
as a basis for the legitimation of their own interests or goals. For dominant
groups, such ideological value-integration obviously will be used to legitimate their domination, and for dominated group to legitimate their opposition, dissidence or resistance (see Chapter 26). That is, the fundamental
legitimacy of any ideological group presupposes that it remains part of the
cultural moral order. Few racists openly defend inequality (see, however,
Chapter 28), but will self-present themselves as emphasizing the relevance
of nationalism and their own freedom (from being 'mixed' with others).
When seen by many others as flouting the principie of equality, thus, blatant
racists are therefore usually marginalized. Hence the prominent role of
denials of racism: whatever one may have against minorities, it will never be
self-defined as racism. Thus, once a fundamental value (such as equality or
democracy) is generally accepted in a society, such values can no longer be
simply 'rejected' by groups without losing their credibility, respect or
societal legitimacy.
The differential ideological incorporation of values by different groups
also suggests that values, as cognitive representations, are not limited to
non-ambiguous concepts. 'Freedom' thus means something rather different
for a corporate manager than it does for a union representative. The same is
true for most groups and most values, as complex 'big' values such as
'democracy' show. Theoretically, it would therefore probably be more
adequate to speak of value-complexes. Thus the freedom-complex would
feature, for example, the following components of the desirable goal
described with the concept of 'freedom': (1) we can do what we want to do;
(2) nobody is limiting our actions, etc.
Values are not merely integrated into ideologies, but govern social beliefs
more generally. Also group attitudes of specific social domains may use
values as benchmarks for evaluation, justification and legitimation. For
instance, one of the evaluative arguments used in the rejection of immigration is that the country is l'uñ . For rather fundamental social, cultural (and
probably biological) reasons, 'overpopulation' (and, implicitly, ethnic mixing) are here used as negative values in the application of xenophobic
ideologies to the domain of immigration.
Mental Models
From the social to the personal
One fundamental lack of afi traditional and contemporary approaches to the
theory of ideology is that they do not account for the relation between the
social and the personal in the accomplishment of social practices. We have
seen that ideologies, like knowledge, attitudes and values are social representations, shared by the members of groups. At the same time, each serious
theory of ideology needs to describe and explain how such social representations are constructed and used by individual group members in and by their
social practices in general, and their discourse in particular.
We also know from both research and experience that these social
practices of individuals are not always'in finé with group ideologies. Apart
from the variable constraints of context, there are personal idiosyncrasies,
different personal histories and different personal experiences, among many
other factors that may affect the variable'expressioñ of ideologies in action
by members as individuals. An empirical theory of ideology that would
systematically describe and explain ideological practices also needs to
account for such differences, variation, dissidence and contradictions.
Though ideologies are shared with others, people make individual use of
them, as they do with their knowledge of the language or the attitudes
of their group or culture. Since also these personal and contextual uses and
variation have general properties, they need to be part of a theory of
ideology. In other words, such a theory must describe and explain also how
ideologies are actually used or applied.
Such a theory at the same time explains the opposite process, namely,
how ideologies are gradually acquired, developed and changed in and by
situated social practices, and especially by discourse. Since social beliefs are
not innate, we must assume that they are gradually acquired by social
perception, interaction, and especially in communicative events. However,
these specific events vary individually and contextually, so we have the
problem of how a'unifled' group ideology may develop from such highly
variable experiences and practices. Apparently, a process of normalization
and unification is at work, that enables general, abstract beliefs to be shared
by many or most members of groups, again much in the same way as natural
languages are learned by language users interacting with each other in many
different situations.
Mental models
79
In the discussion of social representations in general, I have already
suggested that besides the social and abstract account of such representations, we need to realize that although they are shared at the group level, this
does not mean that all group members have identical'copies' of the
representations. Rather, we must assume that because of obvious individual
differences of 'ideological socializatioñ in the group, each member has her
or bis own personal 'version' of the ideology. Obviously, this personal
version must be close enough to the abstract group ideology for members to
be able to function appropriately as competent group members. Again,
comparison with the social and shared nature of grammars and individual
knowledge of a language is instructive here.
It should be stressed that personal 'versions' of ideologies are still to be
seen as social representations. In the memory theory used here this means
that such personal versions of ideologies are part of social memory, and not
of personal (episodic). memory. Despite the idiosyncratic nature of some of
the features of this personal version of ideologies (mostly they will be less
complete than the group-level ideology), their overall forro is general and
abstract, and largely socially shared. In that respect they should be clearly
distinguished from the individual uses of ideologies in specific contexts as a
basis for individual social practices and discourse. It is this last aspect of the
relation between ideology and its manifestation in social practices that is the
topic of this chapter.
Mental models
Theoretically, therefore, what we need is an interface between socially
shared representations and personal practices, that is, a theoretical device
that enables us to connect social (semantic) memory with personal (episodic)
memory and their respective representations. Since the early 1980s, cognitive psychology has with considerable .theoretical and empirical success
developed such a theoretical construct, namely, that of a mental model.'
Mental models are representations in personal memory of events or, as the
terco 'episodic memory' suggests, of episodes. Thus, when witnessing,
participating in or hearing/reading about a car accident, people construct a
model of such an event. Obviously, this model is subjective: it represents the
personal experience and interpretation of the event by the participant. Thus,
what people know personally about such an event, as well as their perspective on and opinion about the event, is represented in their subjective,
individual models of the event. For discourse this means that the model is
being constructed for the event the discourse is about.
In a theory of discourse production and comprehension, to which I shall
turn later, the notion of a model is especially attractive, since it accounts for
the (personal, subjective) interpretation of the discourse by language users.
Indeed, we may now simply say that to understand a discourse ultimately
(and via a number of complex processes) consists in the construction of a
80
Cognition
model. Conversely, in discourse production, the model is precisely the
starting point for text and talk: it is the personal knowledge, experience or
opinion about an event that is being used as 'input' to the discourse
production processes. That is, models also account for the traditional notion
of intention and plan. This means that because they are more or less
independent with respect to discourse meanings, models also explain personal variation and biases of discourse or their interpretation. As we know
intuitively, we may construct an interpretation of the text that may be partly
at odds with the meaning of the text, or indeed with that of the intentions of
the speaker or writer.
As suggested, models are essentially personal and subjective. They
embody personal interpretations and experiences of actions, events and
discourse about such episodes, and this is true for all social practices. This
personal dimension may be the result of earlier experiences (old models that
are being activated or updated) that constitute the personal history of each
person, as well as other, more general or abstract personal representations
(personality, personal opinions, and so on).
People are engaged in the ongoing interpretation of the episodes of their
everyday lives from the moment they wake up until they fall asleep (or lose
consciousness). Such interpretations should be seen as contextually relevant
constructions of such episodes iii mental models stored in episodic memory.
These models also account for the familiar notion of an experience. That is,
it is not the 'real' episodes themselves that play a role in our lives, but rather
their personal interpretation or construction as models, that is the way
episodes are experienced. We may therefore call this particular class of
models experience models. It comes as no surprise that such experience
models are built around the central category of self, which gives the
orientation and the perspective to the model, and which defines the essentially subjective nature of experience models. This means that also representations of our future actions, that is, plan, are a type of experience
models, although these will generally be less specific than the ways we
represent actually %ved' experiences.
Besides the subjectivity of everyday understanding of our environment,
experience models finally also account for the notion of consciousness.
Being conscious means (among other things) that we are aware of ourselves
as well as our environment, and are actively constructing interpretations of
ourselves and that environment. 2
People do not only build models of episodes in which they are engaged
themselves, but also models of those episodes they witness and especially
those they hear and read about. In order to be able to distinguish them from
episodic models about personal experiences, I shall here call them description models, in order to emphasize that we know the episodes through
(discursive) description. Since we may also talk about our personal experiences, and typically do so in personal stories, the sets of experience and
description models obviously overlap. Other episodes we only know about
vicariously, that is through discourse, typically so for many of the episodes
Mental models
81
reported in the mass media. As suggested aboye, description models (earlier
called'situation models'), are needed as a basis for text production and
comprehension. It is plausible, however, that description models are shaped
after our experience models, because we tend to understand unfamiliar
episodes in light of those we know personally.
As is the case for all models, also models of events talked or written about
feature specifications of more general knowledge about such events. Thus, a
model of an event during the civil war in Bosnia is not just built from the
unique, specific and new information we get from the media, but also from
an 'application' of general knowledge about civil wars, armies, killing,
ethnic relations and Yugoslavia. It is in this way that social representations
are 'concretized' in models, and social memory related to episodic memory
and subjective representations. As we shall see later, such episodic models
that interpret discourse, will be fundamental in relating ideologies to
discourse structures.
To avoid terminological confusion, 1 shall henceforth use the term
episodic model (or mental model or simply model) to denote any kind of
model in episodic memory, that is, a subjective representation of an episode.
As explained aboye, I use the term experience model (or simply'experience') for those episodic models that represent personal participation in or
observation of episodes in our own lives. More generally the terco event
model will be used to denote any kind of model that interprets events or
situations (personal or otherwise) referred to by discourse. I also make this
distinction here because the current psychological literature rather confusingly deals with different kinds of models, without explicitly distinguishing
them, and because the various notions of model will be needed below in
showing how ideology monitors social practices. Note finally that all models
may represent both small actions and events (like eating an apple), compound or sequences of events (like meetings) as well as large and complex
episodes such as vacations or civil wars. In other words, episodic memory
consists of sets and systems of hierarchically organized models. Part of that
system, that of our experience models, defines our autobiographical 'past'.
Event schemata
Although no general theoretical proposal has as yet been made about the
structure of these models, we may assume that these structures must be able
to manage effectively the interpretation of events, a process that people are
engaged in many hundreds of times a day. This suggests that also here a
handy schema might be at work, or a number of categories and roles or
strategies to construct such pattems for each situation.
Such a schema is hardly obscure, and has been proposed in different
guises, for instance in the theory of social episodes, 3 as well as in the
functional semantics of propositions. Since people not only represent events
in models, but also routinely talk about them, for example, in everyday
stories, it is not surprising that the categories of these models somehow also
82
Cognition
appear in grammatical and discursive structures: setting (location, time),
circumstances, participants (and their various action roles: agent, patient,
experiencer, object, etc.), and flnally an action or event. 4 Indeed, we may
argue conversely, that once we have introduced the notion of model and its
typical event (or action and situation) schemata, these schematic structures
may be seen as the cognitive basis and explanation of discourse structures.
In other words: the structures that organize the way events are understood
will also influence the ways such events are talked about.
Context models
There is one particular type of event that has a crucial influence on discourse
and its structures, namely, the communicative event or situation in which the
current discourse is being produced and/or received. The mental models of
such communicative events will be called context models. Since context
models represent part of our personal experiences, namely, the one in which
we are engaged when communicating, context models are merely a specific
type of experience models. That is, also context models are personal,
subjective and possibly biased, and hence represent the personally variable
interpretations and opinions of communicative events. We routinely tel
stories about them by later activating such context models.
Context models also have the same structure as experience models,
namely, that of a setting (time, location), circumstances, participants and
their various roles, and finally a communicative action (see Chapter 22 for
details). Crucial in context models is the participant category of self as a
speaker/writer or hearer/reader. It is Chis self-category that defines the
subjectivity of discourse, monitors perspective and point of view, and
organizes many other subjective features of text and talk. 5
Context models are also special because they act in turra as the interface
between event models and discourse. They tell the speech participants who
they are, as what they participate in this event, and a lot of other relevant
information and opinions about the present social situation of talk or text.
The pragmatic constraints that influence discourse meaning and form are
represented in this context model, such as the conditions for speech acts
(usually knowledge about what I and my interlocutor know, want or do),
conditions of politeness (such as social status or power), institutional
circumstances, group membership, mutual knowledge, opinions about each
other, as well as the goals and intentions of the communicative event, and so
on.
That is, instead of merely abstractly spelling out these many 'pragmatic'
constraints of contexts, or vaguely referring to the role of context, we now
have a rather concrete proposal for a more explicit cognitive representation
of such contexts. At the sarne time, context models explain how our personal
knowledge about people, actions, events or situations, as represented in
event models, will be expressed in discourse as a function of the information
Mental models
83
in this context model. That is, context models also operate as the crucial, but
hitherto theoretically elusive, control system in discourse processing.
Whereas our knowledge about an event, as represented in experience or
event models, may be relatively stable across contexts, context models
typically represent the changing, ongoing nature of text production/
comprehension and especially of face-to-face talk. Participants continually
update and change their interpretation of the current situation and represent
this in their context models, which in turra will send their information to the
system of (linguistic) formulation or interpretation. Conversely, during the
interpretation of discourse, our context models (including for instance our
assumptions about the credibility of the writer or speaker) may of course
affect the way we represent the events talked or written about, that is, our
event models. In that respect, event models are not only a function of more
general knowledge of the world, but also a function of the mental representation of the context in which they were constructed — the same story in a
tabloid may be interpreted in a different way (be assigned a different event
model) than when it is published in a serious broadsheet newspaper.
Thus, whereas event models may be described as the basis for the
semantics of discourse, context models are the basis for their pragmatics,
that is, their speech acts, politeness moves, variable lexical or syntactic style,
rhetorical figures of persuasion, and any other feature of discourse that
signals or 'indexes' part of the context, such as choice of dialect or sociolect,
pronunciation, formality or informality, familiarity and intimacy, both in
intonation as well as in lexical selection, and so on. In sum, all properties of
discourse that are contextually variable are by definition monitored by these
'pragmatic' contextanodels. (Por simplicity, I use the notion of'pragmatic'
here in the broad sense it has in much contemporary work in this area,
although I personally favour a stricter use of the terco, namely, as applying
only to the speech act or illocutionary dimensions of discourse; see Chapter
21 for details.)
Linldng the social and the personal
We have now construed one side of the interface that links ideologies to
concrete social practices and discourse, namely, the ways individual social
members represent events, actions or situations in models, and how they
manifest, enact or accomplish these in actual acts and discourse. Details of
the (psycholinguistic) processes involved in the'formulatioñ of model
information in words, phrases, sentences and texts, or conversely in the
interpretation of these verbal structures in tercos of underlying models, will
be ignored here. They are beyond the scope of this book, but I shall later
have to say some more about them when I discuss the ways ideologies are
expressed in discourse.
The next step in the theory is to link individual models with social
representations, because this is the important barrier we need to cross,
84
Cognition
namely, how to get from the personal or individual to the shared and social,
and N'ice versa. The enormous advantage of a cognitive component in a
theory of ideology (and the same is true for a theory of discourse and social
interaction) is that this missing link can be defined (also) in cognitive
terms.
In this case, Chis relationship is established by the fact that models
obviously not only consist of purely personal and individual beliefs, but also
of situated instances of social beliefs. Thus, when being involved in a car
accident we not only know about our personal experiences, or about the
colour or the make of our car, and the unique circumstances of this accident,
but in order to construct the model, we also need socially shared knowledge
about cars, accidents, roads, and so forth, in general. In other words, relevant
elements of social representations, such as scripts, will be activated and
instantiated into knowledge that fits the present model. For instance, general
knowledge that cars usually have four wheels may become relevant to
construct that this car also has four wheels, and so on. Such general
knowledge may of course be adapted to the unique circumstances represented in a model (e.g. to represent cars with three wheels). The same is true
for the construction of context models: we need general knowledge about
people, speakers, communicative events, discourse genres, politeness or
social relations in order to be able to construct an adequate model of the
present communicative event.
Note that this relationship of instantiation and contextual'applicatioñ and
adaptation between social representations and models also may be defined in
the other direction, and thus explain the very acquisition and change of
knowledge, attitudes and ideologies. That is, once constructed for specific
events, these models may be abstracted from and generalized, and thus be
transformed into scripts or other structures of socially shared representations. Formally, this process consists in the change from constants into
variables in the propositions that represent the beliefs in the models and the
social representations.
More empirically, this process may be described as follows: having
repeatedly observed or read or heard about specific events, social members
are able to make generalized inferences, and thus construct beliefs that are
relevant for many different situations, so that the beliefs become useful for
their social status as socially shared knowledge. 6
Such inference processes need not even be accomplished only mentally:
discourse itself is capable of making such generalized, abstract statements.
Social members thus exhibit and at the same time practice their ability to
switch from unique, personal representations of event tokens, to the socially
shared, general representations of event types. This also means that social
learning need not be limited to the 'empirical' generalization and abstraction
from experiences, that is, from models.' People may also acquire social
representations directly, by interpreting generic or abstract sentences and
discourses, as is typically the case in pedagogical or explanatory text and
talk. Also, social members already have vast prior knowledge, and may use
Mental models
85
this directly by making inferences that may produce new knowledge from
existing social knowledge.
Evaluative beliefs
Models not only feature unique personal knowledge about events, but also
opinions about them. When observing, participating in, or reading about a
car accident, people may at the same time construct evaluative beliefs about
the (other) driver or about the ('terrible) accident as a whole. These
opinions will become a natural part of the model, as also when we read
about'ethnic cleansing' in Bosnia, or about our interlocutor in a conversation. Hence, both event models of discourse, and context models, feature
personal opinions about the people, objects or events represented in the
model.
As is true for personal and social knowledge, also these opinions need not
merely be personal. Also evaluative beliefs may instantiate socially shared
beliefs, that is attitudes, for instance about car accidenta, traffic or civil wars.
The same processes of activation, instantiation and adaptation are at work
here, and again in both directions — Personal opinions may be seen to be
shared by others, and thus are generalized as social beliefs and attitudes. The
acquisition and change of social representations may similarly be based on
the generalization and abstraction of opinions in personal models. 8
Such acquisition need not only be 'empirical' , that is, based on personal
experiences, but may also be directly inferred from generalized opinions in
opinion discourse, for instance in editorials in the newspaper, or group-based
evaluations of other group members in conversations. The most dramatic
example of the lattér process is the familiar acquisition of prejudices. These
may be based on a single or a few personal experiences that are 'overgeneralized' as general beliefs, or they may directly be derived from
prejudiced propositions in discriminatory text and talle.
Individuals are members of various social groups. If each such group has
an ideology, individuals share in several ideologies at the same time. When
constructing their models, this also means that general beliefs of more than
one ideology may be 'applied' in the model. If these ideologies are mutually
inconsistent, this may give rise to models that seem to be inconsistent. Thus,
a person may interpret or write a news story, observe or participate in a
social event as a woman, as black, as a journalist, as- an American, as a
Christian, as young and as a Democrat, among other identities. The resulting
model may show a unique and seemingly chaotic combination of beliefs
derived from the ideologies with which that person identifies. This is true
both for the models of the events written about, as well as for the context
model that represents the unique communicative event. In many situations,
thus, people will select or prefer one or more of their present social
'identities' to be dominant in the present context model. Thus, the black
86
Cognition
woman joumalist will often, because of professional constraints and expectations, bring to bear, in the current context model of her news writing,
professional attitudes and ideologies rather than her identity as a black
woman. The detailed structure of context models thus provides an explanation for the ways ideologies indirectly 'map' onto communicative events.
The same will be true for the discourse based on such a model, which in
addition may be further constrained and modified by the context model: The
black female joumalist may well have a personal opinion about news events,
but her editor or readers may not approve of her actually mentioning those.
This means that in the later chapters, I need to investigate a large number of
conditions, processes, strategies and contexts that are relevant in the complex expression and accomplishment of ideologies in interaction, text and
talle.
Finally, ideologies and the attitudes based on them not only influence the
formation of contextually variable personal opinions in models, but may also
operate in selective activation of 'old' models (previous experiences), for
instance in storytelling, news reporting or recall of news. One obvious way
in which this happens is that people tend to activate (recall) those models
whose opinions are consistent with those of the group attitudes they share.
Propositions in such attitudes may thus operate as a powerful search cue.
This is well known from emphical work on racism, in which many white
people typically recall negative stories about immigrants, stories that are
consistent with ethnic prejudices. These stories may function as 'evidence'
in prejudiced arguments: 'You read about that in the paper every day.'
Conversely, they may 'forget' or otherwise suppress stories that confirm
negative propositions about their own group. More generally, people may
selectively search for models as 'confirming evidence' in everyday lypothesis testing'
.a
Conclusions
With the introduction of models in a theory of ideology, I now have
established the necessary links between ideologies and the actual social
practices that construct or implement such ideologies, roughly in this order
from 'deep' cultural beliefs, via group beliefs to their manifestations in
social practices (and vice versa) (see Figure 7.1).
Figure 7.1 shows first that ideologies must be based on a system of
cultural common ground, featuring shared general knowledge and attitudes
and their underlying principies, such as values and cultural truth criteria.
Groups select from this cultural base specific beliefs and evaluation criteria,
and construe these, together with other basic principies of their group, as
systems of specific group beliefs that are organized by underlying ideologies. These representations of the social mind monitor the formation of the
social dimension of personal mental models in episodic memory. Models
that are controlled by group beliefs may be called ideologically 'biased'. The
87
Mental models
Social Situation
Discourse
Working Memory
Context Model
Text Representation
Episodic Memory
Event Model
^
Group Knowledge ^
1
Group Altitudes 1
^
Group Ideologies
Social Memory
1
^
Cultural Common Ground
1
Figure 7.1
personal dimension of these mental models is monitored by old mental
models (earlier experiences) as well as by general representations (personal
knowledge, self, personality) of individuals. Finally, under the constraint of
context models, these personal event models and experience models may be
expressed in discourse or enacted in other social practices.
We have also seen that because of the nature of discourse, also shortcuts
are possible. That is, whereas ideologies are theoretically linked to discourse
only in the indirect way described aboye, discourse may also directly
express fragments of attitudes and ideologies. And vice versa, ideologies are
not merely learned and changed because of personal experiences, but may
also be constructed, at least partially, directly from ideological statements in
discourse. Political and religious conversions are sometimes of this nature.
Propaganda precisely has the function of directly affecting the attitudes and
ideologies of social members, even when 'examples' or'fflustrations' may
be given as persuasive 'evidence' for the validity of general beliefs. This
possibility of a direct link between discourse and ideology also explains the
familiar strategies of manipulation, as well as the classical notion of false
consciousness. Since discourse need not be limited to the expression of
88
Cognition
personal experiences, and hence to concrete social and economic conditions
of social members, ideologies may also be acquired more directly, that is,
through argumentation and other persuasive means to communicate ideological beliefs.
However, most ideologies that control everyday life are gradually
acquired on the basis of a large number of personal experiences and
discourses, and hence do have their 'empirical roots' in personal models. It
may be assumed that such ideologies are also less easy to manipulate,
because they need to be consistent with prevalent experience models.
However, for all situations where social members have fewer, biased or
incomplete personal experiences (models), it will be much easier to manufacture ideologies that have no 'grounding', but which members acquire as a
result of propaganda by elites who control the means of public discourse. I
shall examine these and other social conditions of the acquisition and change
of ideologies and their relations with discourse in later chapters. Important
for now is that we have the theoretical instruments to describe such
processes, and especially to analyse what 'goes on' between social practices,
discourse and ideologies.
Another essential implication of the mental-model theory presented here
is that it explains the fundamental aspect of situational or contextual
variation, and hence the possibility of change. Ideologies and other social
representations are general and abstract, and more or less permanent.
However, we have also witnessed that in specific actions and discourse,
there may be considerable personal and contextual variation in the expression or 'uses' of ideologies. Indeed, because of these considerable personal
variations, empirical studies of ideology (especially in political science)
sometimes conclude that there are no general group ideologies at all.
In the present framework, we are able to account both for the frequent
observation that many group members in many situations do act and tafic
more or less in the same way, while on the other hand accounting for the
uniqueness of all individual actions and discourse, as it is based on personal
models. Since models incorporate instances of social beliefs, while also
featuring personal knowledge and opinions, their expression in discourse
and action may very well have the chaotic and contradictory nature that is so
often observed, in discourse studies as well as in social or political surveys.'
Moreover, individuals are members of various social groups, each with their
own ideology, and as individuals they may, depending on context constraints, draw on several ideologies at the same time, thereby also possibly
exhibiting contradictions that express conflicting interests between these
groups. I shall return to this issue of variation and consistency in the next
chapter.
Thus, whereas ideologies are themselves the interface between the 'social
mind' shared by group members, on the one hand, and social structure, on
the other hand, models are in turra the interface between the social and the
individual, and hence between the general and the particular, and between
shared representations and the actual practices that generate or manifest
Mental models
89
them in concrete social and personal situations. Without this last interface,
we are unable to describe the cognitive basis of (unique) social practices and
discourse, or (with the exception of discourse of a generic nature) to explain
how these are monitored by ideologies.
Consistency
Consistency versus variation
One problem that has often come up in discussions about ideology is
whether ideological beliefs forro a consistent system. Both traditional
work in political psychology as well as in discursive and rhetorical psychology suggests that ideologies are hardly consistent. Both in their actions
and in their text and talk, people show many inconsistencies and dilemmas,
and these do not seem to presuppose neatly consistent underlying
systems.'
The problem with these observations is that while they are undeniably
correct, they do not allow firm conclusions about the structures or the
contents of ideologies. This is not only true because such studies rarely have
an explicit concept of ideological structures in the first place, but rather
because they confuse situational, contextually bound expressions or uses of
ideologies with the ideologies themselves.
We may compare this inconsistency argument with that in linguistics
about the role of grammar. If we examine peoplé s spontaneous talk, we
may observe that they do not exactly always follow the roles of abstract
sentence grammars. To conclude from such personal, contextually variable
uses of a language system that there is no grammar (or that the grammar is
incoherent) is of course hardly persuasive.
Now, if we apply the same argument to the domain of ideologies, where
people not only 'follow' one but possibly many different ideologies,
depending on their various group memberships, the conclusion is even less
persuasive. That is, we not only need to account for such variation, but also,
and perhaps more importantly for the many situations in which social
members do follow their ideological orientation. That is, ideologies should
not be studied merely for isolated contexts or single group members, but
across contexts and for many group members. If such comparisons never or
seldom allow any ideological continuity, then we would have to give up the
very notion of ideology as a monitoring underlying system. In that case we
do have to explain, though, how social members are able to interpret the
various social practices in terms of what usually are called ideologies, such
as 'He is a conservativé , 'He is a racist', 'She is a feminist', or 'She is an
environmentalist', and so on.
Consistency
91
Coherence and consistency
Although we may argue that variable ideological expression does not, as
such, entail the inconsistency of underlying ideologies, the opposite conclusion need not be true either. Forms of ideological continuity of social
practices do not as such imply ideological consistency, at least not in a strict
logical sense. At the very least, we need to assume that if there is some
'order' in ideologies they are at most psycho-logically consistent. Indeed,
ideologies are cognitively and socially constructed, naive basic 'theories' of
social life, and especially about groups and their relations to other groups.
That many members will only acquire and use sometimes incomplete and
inconsistent fragmenta of such ideologies seems obvious, and has already
been accounted for (see also below).
What about the ideology at the social level, however — that is, as a shared
ideology of a group? Again, the comparison with the acquisition of
grammars is instructive here. There is no doubt that people learn their
mother tongue in highly disparate, variable social contexts, and observing
the language use of many (especially less educated) members, one might
conclude that their grammar is hardly complete or perfect. Yet, for their
daily lives they manage quite well in communicating.
The same is true, at the level of individual members, for the sometimes
fragmentary ideologies members acquire as a result of discourse and social
interaction. But at a more abstract level, grammars and other forms of
knowledge should also be considered at the level of a whole community.
Some people in the community know the 'official' grammar better than
others, and formal education precisely involves the teaching of such grammars to the young. Ideological teaching may not, as such, be a subject in
most schools, but both in socialization as well as in formal education and the
later uses of mass media and everyday conversations or other discourses,
there are many examples of inculcation or 'ideological learning'. There are
many situations in which members are able to compare their experiences
(models), including their opinions, as well as their attitudes, with those of
other social members.
That is, the evaluative system of individual members gets 'normalized'
with respect to the social beliefs of the group, community and culture as a
whole. As is the case for grammar, people may well not have active
knowledge about the precise contents or structures of such ideologies, but, in
the evaluation of their own social practices and those of other social
members, they should in general be quite competent in making the 'right'
evaluations, and follow the ideological principles that are in the interests of
the group. This is the case despite the fact that people may be manipulated to
adopt 'inconsistent' ideological principles when they lack adequate information or personal experiences to draw on.
As basic group beliefs, ideologies are not merely based on the experiences
of a few members, but grounded in the socially and historically developed,
accumulated and (discursively) transmitted experiences of the whole group,
92
Cognition
now and in the past. Such collective experiences will be a powerful
corrective to fully inconsistent and incomplete ideologies. Intra-group discourse will provide the experiences, the evidence, the argumenta, the
situations, and so on, that are necessary for the whole group to develop its
group ideology, even when individual members do so imperfectly or quite
variably as a result of their specific social position or as a consequence of the
influence of other ideologies. For large, institutionalized or otherwise
organized groups, there will be special ideological institutions (training,
lectures, seminars, media and propaganda) that may spell out the details of
such ideologies, as is the case for religions (Churches), political parties,
unions, non-governmental organizations (like Greenpeace), or large social
movements such as the womeñ s movement, as we shall see in more detall
laten'
The crucial point here is that although not all individual members need to
be able to explicitly formulate the ideologies of the groups of which they are
members, groups as a whole may still develop complex and more or less
coherent group ideologies. Such groups will have leaders or other elite group
members (the ideologues) who know and teach or transmit such an ideology
to new members. If most group members were to have highly fragmentary or
incoherent ideologies, they would be unable to organize their relevant group
attitudes and forro the models that are necessary for their everyday practices
as group members.
In sum, given various social constraints on groups and group relations,
and the collective and historical experiences of their members, we may
therefore provisionally conclude that, at the level of the group, ideologies
should be fairly stable and coherent.
Conditions of variation
Situational and personal variation is easy to explain while maintaining the
notion of an underlying ideology that is more or less coherent. Let us
examine some of the conditions of such variation.
The first reason is that ideologies are not always immediately linked to
discourse, but usually indirectly, that is, via more detailed knowledge,
actitudes and personal episodic models. That is, in their everyday lives,
social members rather operate at the 'meso-level' of group beliefs, rather
than at the high, abstract level of ideologies. For instance, they may be
aware of and apply opinions on immigration or unemployment rather than
abstract ideological principies of a racist (or anti-racist) ideology, although
the latter may sometimes be made explicit in accounts and argumentations,
and appear more often in the discourse of the elites.
Second, ideologies, as well as actitudes and knowledge, are socially
shared, and hence 'context-free' in the sense of being stable across different,
specific social situations at the micro-level. In these everyday contexts and
practices, people deal with more concrete events, people and situations, as
Consistency
93
represented in mental models. These models are strategically adapted to the
situation at hand, and this means that sometimes the expression of an
ideologically'correá opinion may be less appropriate for reasons of
politeness, positive self-presentation and commonsense interests — racist
shopkeepers would soon be broke if they were to openly derogate their black
customers, for instance.
Moreover, sine social members are members of several groups, they'll
bring to bear several ideologies in their models of everyday events, so that
the models may become seemingly incoherent. The same is true for their
practices and discourses. People may adhere both to more or less humanitarian and democratic principies, but at the same time not apply them to
certain social relations: for example, those of gender, age or ethnicity. The
use of various ideologies in one situation (and in laboratory experiments the
same is true for simulated situations) thus results in sometimes complex
models which exhibit apparently incoherent opinions in discourse, typically
expressed in disclaimers such as 'I am not a racist, but .. 'I am ah for
equality of women, but.. .'. Personal experiences and biographies, local
circumstances and interpersonal relations will further contribute to the
complexity of such models and the discourses based on them. Indeed, many
of these observations have already been made, in other terms, in classical
studies on cognitive dissonance, 4 which we are now able to reformulate in
terms of model structures and relations between models and social representations. Hence, as suggested before, conclusions about the contents and
structures of ideologies need to be based on comparisons of many events in
which variable properties of discourse are explained in tercos of such
constraints.
In other words, variability of ideological expression is explained by the
complex interplay of several ideologies and their contextually specific uses,
whereas the continuity of ideological opinions can be explained in terms of
socially shared ideologies that are rather stable and context-free. No need
therefore to assume that ideologies are sets of mutually incoherent propositions. If such were the case, people would in principie always express
themselves incoherently, also across situations, and there is no evidence for
that. On the contrary, we know from work on racism (and sexism, etc.) that
talk on ethnic (or gender) relations rather consistently has recourse, in
variable contexts, to similar basic norms, values, principies, ideological
propositions and more specific attitudes. Unless personal and social circumstances change dramatically, or when being targeted by persuasive discourse, someone advocating liberal immigration rules today, will not be for
strict ones tomorrow. 5
That is, the 'normal' situation is that of individual variability, and the
situation that especially needs to be explained is precisely the fact that many
different people in many different situations still appear to use very similar
ideological opinions. It is ideological conformity and consensus that are
remarkable, and much less that different people with different experiences
have differences of opinion. On the other hand, if such conformity is mainly
94
Cognition
explained in terms of the identical social or economic situation of a group,
then individual variation and dissent are the phenomena that need to be
explained.
Change
Of course, ideologies may change, but this takes time, in particular because
they are socially shared, and groups need much time to change their basic
ideologies because such changes need much public discourse and debate.
And precisely during such periods of ideological formation and change,
other (opposed) ideologies may become more relevant in the control of
action and discourse, and more personal variation will be the result. For
instance, after the demise of state socialism and communism around 1990,
which also affected the Left in general, leftist ideologies entered a state of
transition as well, whereas neo-liberal market ideologies have become not
only more dominant but virtually hegemonic. As a result, even in more or
less progressive media, socialist discourse has become 'unfashionablé . I
shall come back to these social and political conditions of changing
ideologies laten'
Conclusion
From these arguments it may be concluded that ideologies 'ideally' reflect
the goals and the interests of the group, and do so optimally when these
interests are coherently translated into a set of basic beliefs shared within the
group. Such coherence facilitates the organization of new attitudes about
specific issues and the co-ordination of social practices by different members
in different situations. That is, coherence is a condition of continuity and
reproduction. As a theoretical hypothesis it explains members' experiences,
observations and expressions of such continuity.
Variations and contradictions in the enactment or expression of such
ideologies are perfectly compatible with this hypothesis if we assume that
such manifestations are explained by multiple ideological allegiances of
social actors, both at the level of attitudes, as well as at the level of specific,
contextual and personal, models of events, which in tum monitor discourse
and other social practices. As is generally the case for social representations,
including knowledge (including knowledge of the language or about interaction), social members are experts in adapting these shared representations
to their personal needs and contextual constraints.
It is also at this level that systematic variation and 'deviatioñ may give
rise to attitudinal and fmally to ideological change, as soon as enough
members, and especially leaders who control public discourse, are able to
persuasively communicate such alternative systems of judgement to other
group members. Changes in feminist and socialist ideologies are well-known
examples of such transformations. In this way, ideologies, despite their
Consistency
95
relatively stable nature, may with some delay flexibly change as a consequence of (a) changing social interests and (b) the everyday experiences of
group members, and of course (c) persuasive ideological discourse. These
and other conditions for ideological continuity and change will be discussed
in later chapters.
Consciousness
Introduction
When dealing with the cognitive dimension of ideologies, anothér issue
needs to be addressed, namely, that of consciousness. In the first ,place, this
notion has been part of the history of the study of ideologies since Marx and
Engels, mostly in the form of 'false consciousness'. Second, we may ask
whether social group members have, experience or use their ideologies more
or less consciously, or whether these belief systems are acquired, used and
changed more or less 'unconsciously' or, in other tercos,
False consciousness
The traditional notion of 'consciousness' (German: Bewufitsein) plays a
central role in the traditional accounts of ideologies, especially in combination with its negative modifier 'false This phrase then usually refers to
group ideologies that do not reflect the 'objective' socio-economic interests
of a group. Thus, workers or poor people may develop an ideology that is
rather in the interests of, respectively, the ruling class, the elites, the
company they work for, or the owners or the managers of that company.
Such a 'false' or misguided ideology may be the result of a mixture of
ignorance, indifference, manipulation, compliance or concem for short-term
interests (e.g. not to lose oné s job, getting a pay-rise) at the expense of longterm, structural interests, such as ownership of the means of production, or at
least shared planning and decision making The contemporary 'alienation' of
the working class in times of the hegemony of liberal market ideologies is a
well-known example in point: large segments of the working-class no longer
vote socialist nor even social-democratic. Having obtained a minimum of
job security and income, they adopt more or less conservative market
ideologies or forms of liberal individualism.
The social dimensions of this problem will be dealt with later (see Chapter
11). Here, I focus on the cognitive aspect of this classical issue in the theory
of ideology. What, indeed, is consciousness exactly? In the description of
the issue in the previous paragraph, I have freely replaced false consciousness by false ideologies. That is, consciousness is obviously taken as a'state
of mirad', in this case of sets of beliefs. More specifically, since the
'consciousness' of whole groups or classes is involved, these beliefs must be
socially shared. Hence, ,the most adequate translation of the term in my
. 1
Consciousness
97
framework would simply be that of social representation. This comprises not
only basic ideologies of the group, but also their attitudes and knowledge.
The advantage of such a broad definition of this vague term is that not only
value systems and judgements may be involved, but also knowledge.
Indeed, the concept of false consciousness is also used in order to denote
ignorance of the 'real' social facts, for instance about the interests of specific
social arrangements, policies or practices. Again, such ignorance may result
from widespread indifference and apathy, which may again result from
oppression or partial satisfaction with the status quo. Or it may more
actively be inculcated through biased information or by other forms of
ideological manipulation by dominant groups. In that case ideology as false
consciousness of dominated group A in fact implies the acceptance of a
hegemonic ideology of dominant group B, for instance as those beliefs that
misrepresent social inequality as the allegedly natural or immutable nature
of the current social and moral order. In other words, the notion of 'false'
here also implies conceptions about truth and falsity of beliefs and ideologies about social life, to be discussed in Chapter 11.
In sum, my analysis of the traditional notion of 'false consciousness'
proposes to make the notion of 'consciousness' more explicit, first of all, by
using the term 'social representation', including knowledge as well as
(evaluative) attitudes and ideologies. Second, the notion of 'false' may be
assigned two different meanings: (1) wrong, partial, incomplete, biased or
otherwise misguided factual beliefs (which presupposes that there is 'correct' or 'true' knowledge; see Chapter 11), and (2) evaluative beliefs that
lead to judgements and social practices that are not in the interest of the own
group, and may be in the interest of a dominant group. After this conceptual
clarification of a classic notion, I later study in more detall which discursive
conditions and what social situations may bring about such non-self-serving
social representations.
Having somewhat clarified the notion of (false) consciousness, we remain
with the empirical question whether the main (Marxist) theses of ideological
domination and inculcation and the construction of 'misguided' ideologies
by dominated groups is correct in the first place. That is, first of all, there
may not be just one 'dominant' (class) ideology in the first place, but a
complex structure of elite ideologies that may mutually compete for control
or hegemony. 2
Second, although it is plausible that when the ideological elites have
control over the means of ideological production (especially politics, education and the media) their social representations about society may be quite
influential, this does not entail that dominated groups will actually adopt
such representations. Whereas such ideological influence may be plausible
in situations in which no alternative knowledge and opinion sources are
available or accessible, and if dominant attitudes do not obviously clash with
the immediate interests of dominated groups and their members (typically so
in racist ideologies), such inculcation is much less obvious when group
98
Cognition
members are able to observe directly the contradictions between the inculcated ideologies and attitudes and their everyday experiences. Indeed, if
such were the case, social and individual resistance and change would be
difficult or even impossible. These issues will be 3dealt with in more detail
when I study the social dimensions of ideologies.
Consciousness as awareness
A related problem of the definition of (false) consciousness, and hence of
ideologies, is whether people who have' them are actually 'aware' of them.
We already have seen in the previous chapter that this is not necessarily the
case. Social members are barely aware of many of the social representations
they have, and of the ways these control their social practices and evaluations. As we shall see later for the notion of 'common sense', ideologies
may be or seem so 'natural' that people doñ t even realize they have them.
As is the case for knowledge of natural language (people' s shared competence), ideologies often are simply part of everyday life, and taken for
granted. 4
While this is true and probably applies to many ideologies, it is hardly a
property of all ideologies. Some ideologies are explicitly 'invented' in the
appropriate historical and social circumstances and explicitly propagated
among group members. Especially for dominated groups, thus, it is crucial
that the ideology be made and kept conscious, and there are many organizational practices that will make sure that such is the case: media messages,
party or group meetings, socialization events, initiation rites, propaganda,
and so on, will make sure that members leam to be aware of the ideological
basis of their group membership. 5
Unlike much implicit grammatical knowledge, some ideologies may
partially be made explicit in everyday discourse, for example when people
defend their own ideologies or attack others. Parts of the arguments in such
debates will be based on ideological principies that may need to be
formulated explicitly as premises of an argumentation. This means that
language users of the ideological group get frequent 'exercisé in acquiring
the group ideology.
Whereas oppositional ideologies by defmition will tend to be more
explicit and conscious among group members, dominant ideologies will
precisely tend to be implicit and denied, or felt to be 'natural' by their
members. Such group members may indeed be unaware of their ideologies
(typically so of male chauvinism, racism, etc.) until they are challenged by
members of the other group.
Although the notion of 'consciousness' may be clear in it its everyday
sense, it is theoretically hardly explicit. Being conscious of, being aware of
or realizing something, first of all, is a 'state of mirad'. For instance, it may
mean that active processes of thinking, mental arguing or simply information search have access to specific information. If so, people are able to use
such information in arguments, or for further processes of inference.
Consciousness
99
In other words, there are many types or grades of 'consciousness' between
totally explicit awareness and knowledge on the one hand, and largely
implicit knowledge and 'mere' use as %ved experiencé on the other hand.
This distinction is socially often associated with that between 'ideologues'
and the 'masses', a distinction we need to analyse as part of the social
analysis of ideologies.
Cognitively, the distinction means that at least some group members — and
for each group this fraction may be different in size — not only share the
ideology but also explicitly know its major tenets, and are able to talk about
them as such, and even argue for them as such. It has been frequently
observed that such explicit ideological self-awareness is rare, and usually
limited to leaders, opinion-leaders and other elites. These are also those who
have the function to formulate and persuasively inculcate the relevant
ideological beliefs among group members — for example via propaganda — or
who 6are able to explain relevant everyday events in terms of the ideology.
On the other hand, such explicit knowledge of ideological beliefs may be
limited to only a few basic beliefs, or only be accessible in their more
specific form, for instance as opinions in particular attitudes. For instance,
people may be unable to spell out general racist or ethnocentric principies as
such, but they very well know that they disapprove of liberal immigration or
of preferential job and housing allocation to minorities or refugees. That is,
in that situation, their knowledge is still explicit, general and social, and may
be expressed as such: 'We in our community think that .. Another
intermediate stage of awareness is when group members have such attitudes
on some issues, but not on other relevant ones. For instance, they may share
a prejudiced attitude on immigration, but have (as yet) no ethnic attitude on
educatión or language use.
Finally, largely implicit are those ideological opinions that only exist at
the level of models of concrete events: for example when someone does not
want a foreign neighbour, but does not generalize or rationalize this
explicitly in terms of group attitudes ('We doñ t want foreign neighbours
because ..."). Social interests in this case are completely translated and
integrated into personal interests. I surmise, though, that such totally implicit
knowledge and attitudes are rare in most contemporary societies, where most
members have access to the mass media: in such a situation people quickly
learn to legitimate their personal opinions in terms of the shared attitudes of
the group. Systematic discourse analysis of ideological text and talk provides insight into these various leyels or grades of ideological awareness.
Also, I later need to demonstrate whether and how such awareness may be
enhanced (or suppressed) by communication: for example by party or group
propaganda, teaching, seminars, meetings, media and so on. The fact that
many ideological groups organize various forms of 'consciousness raising'
or 'awareness training' suggests that such forms of ideological explication
may be an important organizational feature of ideological groups, especially
for dominated groups or social movements.
100
Cognition
The contemporary uses of the notion of consciousness in cognitive
science are quite different from the traditional meaning of (false) consciousness. In current debates the concept of consciousness applies to the complex
problem of how we can explain how brains can be associated with the (self-)
consciousness of minds. In many respects this problem boils down again to
the eternal problem of brain—mind identity and difference. Once it is
accepted that, at some level of analysis, brains also have mind-like qualities,
we at the same time explain the elusive notion of consciousness — for
example, in tercos of knowledge of self, knowledge about the present
context, and especially in terms of ongoing mental processes like thinking
(including the ability of people to be able to think about themselves and their
own thinking). It is this kind of consciousness that especially characterizes
the ongoing construction of mental models of experience (see Chapter 7).
With our contemporary knowledge about the brain and the mind, there
does not seem to be an additional property of either of them that needs
special explication in terms of consciousness or awareness. An operating
human mind is by definition 'conscious' when people know about themselves, their ongoing activities and their ongoing thoughts. Then there is the
distinction between 'conscious' and 'automatic' processing of information.
This suggests that consciousness and mind cannot simply be identified —
processing in the first case involves self-awareness of short-term memory
processes, or as knowledge of what one is now doing. In Chapter 7 it was
proposed to represent at least part of this knowledge in experience and
context models, which therefore may also be seen as the kind of overall
mental monitor sometimes associated with awareness. Of course, this does
not solve all problems related -to the notion of consciousness, but for my
discussion it should do.'
Awareness and denial
Finally, we need to realize that more-or-less explicit knowledge of ideological beliefs of group members who positively identify with a group usually
implies positive acceptance of such beliefs. This also means positive selfpresentation and description of such beliefs. It is well known for instance
that virtually all racists will deny that they are racists, and many of them
seem to reject racist ideologies when described as such. 8 However, when not
described as racist, but, for example, as nationalist, or as 'normal' or
'natural' beliefs in favour of the own group, then the same ideological
beliefs may well be perfectly acceptable. In other words, knowledge and
acceptance of ideologies and their derived opinions by group members
usually implies acceptance of ideologies-as-described-by-the-group-itself.
The converse is true for the rejection or change of (other) ideologies, as is
the case for anti-communism and anti-racism, which are based on otherdescription of ideologies. Such opposite ideologies may again be reflected in
the argumentative, explanatory or legitimating discourse of group members
Consciousness
101
sharing the ideologies thus criticized, for instance in well-known disclaimers
such as 'I am not a racist, but .. or '1 am not a sexist, but ..', and so on.
Indeed, given the negative meaning of the concept of ideology in everyday
usage, groups and their members may deny that they have an ideology in the
first place. Thus, being in favour of market-freedom will seldom be selfdescribed by its adherents as a belief that characterizes a 'capitalise
ideology, or even as a 'liberal' ideology. Similarly, Christians or Muslims
will not usually self-describe themselves as adherents of religious ideologies. At most, terms such as 'philosophy', 'principies', 'convictions' or
simply'belief(s)' will be acceptable as self-descriptions of ideologies.
I shall later investigate such strategies of self-description, denial and
legitimation in the expression and defence of ideologies in more detall. Por
my present discussion such examples only show that people are aware of
conflicting ideologies, that they know that their expression of specific
opinions may be 'heard as' expressions of a normatively unacceptable
ideology, and that they usually self-represent and defend their own ideologies in positive terms.
10
Common S ense
The meanings of 'common sense'
Related to the notions of consciousness and awareness is another sociocognitive notion that plays a central role in contemporary discussions of
ideology, namely, common sense. This notion has roots in different philosophical and sociological traditions.
First of all, it is often associated with Antonio Gramsci's contributions to
the theory of ideology, and especially with the concept of hegemony.' As
soon as groups and their members accept a dominant ideology as a reflection
of their own goals, desires or interests, or as a representation of a natural or
otherwise legitimate social order, their ideologies may turra into beliefs that
are taken for granted or simply common sense. Ideological dominance and
hegemony is 'perfect' when dominated groups are unable to distinguish
between their own interests and attitudes and those of dominant groups. In
that case, they may not even be able to see conflicting ideologies (even when
in their own best interests) as viable or acceptable alternatives. I shall later
come back to the social dimensions of such forms of ideological compliance.
Another main source of the notion of common sense may be found in
phenomenological microsociology and ethnomethodology, for example as
inspired by Alfred Schütz. 2 Here common sense is simply defined in terms
of the implicit social knowledge that group members take for granted in
their everyday social practices. This members' knowledge is essentially
shared lay-knowledge, and should be distinguished from elite or theoretical
formulations or explications of knowledge. Indeed, compared with explicit,
scientific knowledge, commonsense knowledge may sometimes be described
as wrong, biased, misguided or otherwise unfounded. However, outside such
a critical account of common sense, it should be emphasized that whatever
the truth-status of commonsense beliefs, they are usually true and accepted
by the people who hold them, and will therefore be at the basis of their
everyday experiences and interactions, that is, their practical accomplishments. In other words, for group members they are true 'for all practical
purposes'.
Describing and explaining such mundane practices, therefore, also
requires making explicit the similarly mundane beliefs (methods, rules, etc.)
that group members take for granted. This means that their actions, including
their discourses, will be described from their own point of view, and
Common sense
103
possibiy in tercos of the notions and categories they use themselves. Indeed,
using the theoretical categories of the sociologist may fully misrepresent the
ways members understand and accomplish everyday activities. In other
words, a theoretical account of common sense and what is taken for granted
in interaction, at the same time becomes a methodological principie, namely,
to study social reality as much as possible from the point of view, and in
terms of, the social actors themselves.
One important implication of the notion of taken-for-granted knowledge
for the study of discourse is that such knowledge tends to be presupposed.
That is, such beliefs are not explicitly stated, but incorporated without
challenge in new statements about social reality, because language users
may assume that the recipiente have similar beliefs, and indeed similar,
recognizable 'methods' to organize everyday interaction in general, and
conversation in particular. This link between common sense, knowledge and
discourse will be explored in more detall later (Chapter 11).
For my discussion, these various notions of common sense, and especially
the Gramscian and the ethnomethodological ones, also suggest elements for
a theory of ideology. For this reason, contemporary studies of ideologies
tend to emphasize the implicit, taken-for-granted, common-sense nature of
ideologies as %ved experiences' in the everyday lives of groups and their
members. 3 In light of the discussion about consciousness and awareness in
the previous chapter, this conception of ideology identifies ideologies with
the non-conscious, unaware mode of ideological practices. People simply go
about their everyday business and spontaneously see and judge social reality
and events in terms of a belief system that is normal and unproblematical,
and which they assume is shared by other group members. Only in situations
of complications, challenges or other deviations from the accepted system of
knowledge, may group members be (made) aware of the problematic nature
of their commonsense or ideological beliefs. In such situations, however,
they may have similarly commonsense 'methods' to deal with problems and
try to resolve them for the situation at hand.
What is common sense?
Given my earlier discussion, this account of common sense and ideology
explains only part of the facts. Where the notion of common sense is
relevant, I first of all need to make it explicit. As so many 'mental' terms in
the philosophy of ideology, and the microsociology of everyday life, this
notion was cantil recently seldom made explicit beyond a characterization in
tercos of mundane, taken-for-granted beliefs. But we have seen that there are
many kinds of belief in the realm of cognition or memory, so I need to
specify which ones can be seen as commonsense beliefs.
Our proposal for definition will again be straightforward — common sense
is just another term for the set of social beliefs. Like the latter, it is social,
shared by members of a group or community, and involves knowledge as
104
Cognition
well as opinions. 4 In those respects, common sense is a modern variant of
the notion of consciousness discussed in the previous chapter.
One dimension of common sense, however, absent in the intuitive notion
of consciousness is its argumentative nature. Common sense is typically
referred to, especially in everyday (commonsense!) uses of the term, when
arguments are said to be based on common sense.
In other words, un]ike sociologists, who take common sense as the shared
knowledge underlying all mundane interaction, the commonsense meaning
of common sense usually involves discourse: arguments, accounts, explanations, defences and legitimation. 5 More specifically, it connotes that such
knowledge is direct, immediate, unreflected, untheoretical and unscientific,
but based on or derived from everyday observation or experiences. Common
sense in this sense is an implicit, naive 'theory' of the world. 6
More critically, this same explanation may therefore imply that common
sense is essentially unreliable, possibly biased by social prejudices and
iflusions, if not the result of manipulation. In both cases, common sense is
also associated, at least implicitly, with popular or lower-class cognition, as
what 'common people' think and find.
We see that common sense has many meanings and various critical
interpretations, which need analytically be kept apart. Hence, we first keep
its basic meaning, that is, as social representations, in order to account for
taken-for-granted knowledge, but add that this may also include taken-forgranted other beliefs, such as socially shared opinions (and prejudices).
They take away our jobs', is a typical example of such commonsense,
prejudiced opinions.
Next, the argumentative and discursive role of common sense needs to be
represented differently, in terms of the ways social representations (knowledge and attitudes) are activated, used and contextually adapted in communicative events, that is, as part of specific models. A commonsense
argument, then, is an argument based on a common sense model, that is, a
model of which much of the knowledge and opinions is largely shared by
others. The same is true for commonsense descriptions, accounts and
explanations. Such accounts are typically founded on 'what we all know on
'what everybody says', (consensus), or on commonsense truth criteria
have seen it myself ).
Third, the dimension of common sense as being immediate, unreflected
and untheoretical may simply be described in terms of the type of social
representations being shared within a group (expert versus lay knowledge,
etc.), and also in terms of the more or less unproblematic processing of
social knowledge. Models are directly formed from instantations of general,
shared knowledge, and not by independent, critical examination of the
'facts', nor by more complicated thinking or reasoning. Hence also the elitist
association of common sense with what is taken-for-granted by the uneducated 'masses'. This need not always be a negative implication: common
sense is also positively valued as an antidote against scientific sophistication,
jargon, and needlessly complex explications of what 'ordinary' common
Common sense
105
sense will tell us more directly and more transparently. In Chis sense,
common sense positively reflects what is 'obvious' and 'for all to see',
against the pretensions of fancy scholarship. When used in this sense, it may
also be a tenet of anti-intellectualism. 7
common
And finally, Chis sociological dimension — who indeed
sense, and who uses it — needs to be described on the basis of a sociology of
knowledge and a study of discourses (and especially argumentations and
explanations) by various groups and their members. In sum, a theory of
common sense examines its structures and status as social representations,
its processes or strategies in thinking and its uses in social practices and
discourse, and its uses by specific social groups.
In this cognitive section, we may for instance examine in which respect
common sense, as representad in social representations, is being used in the
formation of models: in interpretations of events, as personal instantiations
of social knowledge and other beliefs, as being strategic (fast but not perfect)
and, as we shall see later, as being largely implicit, that is, as not explicitly
commented upon in discourse. 8 One, more romantic association of common
sense as mode of thought, may be put to rest from the start: in many
contemporary, mass-mediated societies with virtual universal literacy and
bigh education levels, there is hardly such a thing as 'pure' common sense,
in the sense of shared, unreflected, untheoretical knowledge, based only on
our experiences. Precisely one of the reasons why Serge Moscovici and
French social psychology introduced the very notion of social representations was to emphasize the 'popular' integration of scientific theories. 9 The
best known case of this phenomenon is the now common use of notions
from psychoanalytical theories.
Similarly, elites who have special access to the media, and hence
indirectly to the minds of the public at large, will routinely describe and
explain events in terms of implicit or explicit scholarly theories, and this will
obviously also influence the social representations and explanations of other
group members. It follows that although everyday perception and anderstanding may well be based on personal experiences and on a more or less
unreflected application of commonsense knowledge in the construction of
models, these socially shared representations also involve more or less
simplified versions of scholarly knowledge. The same is true for truth
criteria, inferences and argumentation. Accounts and explanations have
become largely acceptable only when based on truth criteria that are
themselves socially and culturally variable versions of more philosophical or
scholarly ways of arguing and thinking Asking the opinion of a sorcerer,
examining entrails or the fines of oné s hand, or looking at the stars, among
many other remnants of old popular criteria of truth, have been largely
discredited as superstition. In sum, in most modem societies, there is no
'pure and popular', scientifically uncontaminated, common sense, but rather
a gradual difference with explicit, scientific, methods of observation, thinking, proof and truth criteria.
106
Cognition
More generally, we may conclude that the difference between group
beliefs and cultural beliefs is relevant for a theory of common sense, and
most of what we have said aboye also applies here. That is, specific group
knowledges and opinions may slowly be integrated in (or excluded from) the
cultural common ground. Common sense is then more or less what we try to
conceptualize with the term 'cultural beliefs', that is, the knowledge and
opinions, as well as the evaluation criteria, that are common to all or most
members of a culture. Like common sense, these cultural beliefs are also
used as the basis for specific group beliefs, and also function as the general
base of presupposed beliefs in all accounts, explanations and arguments.
Ideologies as common sense
The same applies to the identification of ideologies as forms of common
sense. Depending on context and social group, ideologies may be more or
less known and used explicitly in the conduct of everyday life. Thus, we
may distinguish between the explicit treatises of the 'ideologues' and the
ideological 'commonsensé reasoning of other group members, but should
be aware that these different modes of thinking and discourse mutually
influence each other. History has shown that much of what once counted as
'scientific knowledge' (e.g. about women or blacks) may now be rejected as
unfounded, if not prejudiced 'common sense'. 1.
That many everyday actions are being accomplished routinely, and
seemingly unreflectedly, does not mean that members are unable to make
explicit at least some of the knowledge and other beliefs that are implied or
presupposed by their practices and discourses. Misunderstanding, conflicts,
challenges and various factors of the context may give rise to various modes
of 'explication', in the double sense of the term: as making explicit, and as
explanation or account.
Both the social representations on which such explications are based, as
well as the nature of the explications themselves, namely, as valid and
acceptable arguments, may be more or less explicit and more or less imbued
with widely shared, popular versions of scientific knowledge. This may be
true more often and more explicitly among members of specific (elite)
groups, but my point is that because of general education and the media,
such philosophical and scientific influences on 'common sense' may be
fairly widespread among many ideological groups.
Thus, most members of environmental groups have a fair amount of more
or less technical knowledge about the nature, the causes and the consequences of pollution. Feminists may have extensive knowledge and
attitudes about gender relations, and their arguments may be based not only
on the shared immediate experiences of all or most women, but also on
scholarly research or intellectual argument.
Concluding we should emphasize that if common sense is identified with
the general beliefs of a culture, and if ideologies as the foundation of
Common sense
107
specific group beliefs are based on such a cultural common ground,
ideologies themselves are not aforro of common sense. Indeed, the very
'common' in common sense implies that such 'sense' is being shared, and
hence rather cultural than group bound. Moreover, ideologies are usually
much less taken for granted than are general cultural beliefs, because they
are often more explicitly taught within the group and contested and hence
defended across group boundaries. People are usually more explicitly and
consciously Christian, socialist or feminist than 'Western'. Only in crosscultural conflicts are people made aware of the common sense of their own
culture. In other words, ideologies as defined here should not typically be
identified with common sense, but rather with uncommon sense or nonsense.
11
Knowledge and Truth
Ideology versus knowledge
In many classical approaches as well as in most commonsense and political
conceptions, ideologies are typically described as false, wrong, misguided,
and as such opposed to true — and especially scientific — knowledge. Full
discussion of the issues involved here would require a monograph by itself.
So I shall only briefly summarize some major tenets and take a position that
fits the theory presented in this book, elaborating the suggestions made in
Chapter 3. 1
The critical opposition of ideology and knowledge goes back at least to
Marx and Engels and their conception of 'false consciousness', which
implies that in specific situations and under the influence of ruling class
manipulation, the working class may have misguided beliefs about the
material conditions of its existence. Dominant ideologies in that case are an
instrument of the ruling class which serves to conceal its power and the real
socio-economic conditions of the working class. Throughout the history of
political economy and sociology, similar distinctions were made, usually
opposing ideology to scientific knowledge, that is, (with Durkheim) the
'sociological facts' established by social science. Until today, as we have
seen before, ideology is thus characterized in tercos of common sense, as
beliefs that are taken for granted, and in general with naive views of
everyday life that may be at variance with the knowledge produced by
objectivé scholarship.
It is hardly surprising that these views have also met with considerable
critique. Thus, it has been pointed out that the history of science clearly
shows how much scientific knowledge and methods themselves may be
based on ideologies that are in the interest of the elites, if only in the interest
of scholars themselves. From a different, ethnomethodological, point of
view, commonsense knowledge of social members has received a more
positive evaluation in tercos of the practical basis of social practices, and as
a viable means by which members manage their everyday lives. 2
Against the background of this briefly summarized history of the opposition between ideology and knowledge, we should finally examine the role of
knowledge in the conception of ideology presented in the preceding chapters. It was assumed that ideologies form the 'axiomatic' base of the social
beliefs of a group. These social beliefs may be factual or evaluative. Por the
evaluative beliefs (opinions, actitudes) of a group, which may be typically
Knowledge and truth
109
contested by other groups, it is rather easy to accept that they are ideological.
But what about their knowledge? How can knowledge be ideological, and
still be called 'knowledge', that is, 'true belief , in the first place, instead of
being characterized as 'mere' group belief (in the everyday sense of that
term) or as 'opinioñ . Or should we as sume that, since all knowledge criteria
are historically and culturally variable, also all knowledge is relative and
hence possibly 'ideological'? Let us examine these questions in more detall,
and reformulate some tentative answers within the theoretical framework
presented in this book.
Por various theoretical reasons, it was assumed that ideologies essentially
involve values and therefore monitor the evaluative beliefs of groups, that is,
attitudes. One question that may be raised in that case is whether ideologies
may also influence non-evaluative, factual knowledge, or even whether more
generally we should adopt the view that all knowledge is ultirnately
ideologically based. We might call this the ideological relativism thesis,
following the more general view that all knowledge is socially and culturally
relative, given the historically and culturally variable nature of truth criteria
that forro the basis of such knowledge. Let us examine whether this thesis
can be defended within the framework of this book.
The nature of knowledge
Both in everyday life as well as in epistemology, knowledge is usually
defined as justified true belief. Thus, in common sense language use, we may
adequately say that we know that p if we believe that p and if we have good
reasons, evidence or proof thatp is true. That is, if called finto question,
knowledge statements may have to be justified, for example in terms of
culturally accepted truth criteria, such as personal observation, reliable
sources (media, experts, etc.), logical inference, common sense or consensus (Tverybody knows that ...D. Similarly, again in everyday
discourse, we attribute knowledge to others, rather than mere beliefs, if what
others believe is true according to us, that is, if someone else shares our
knowledge. On the other hand, we use the word 'belief to denote those of
our own beliefs for which we have no, or insufficient, evidence, or those of
others which we know to be false or about which we have insufficient
evidence.
Episternology provides further conditions for (rather marginal) cases of
(lack of) justification, for example when someone believes something that
happens to be true, but has the wrong (non-justified) reasons for doing so. I
woñ t go finto these and other complications of the contemporary philosophy
of knowledge. Similarly, I shall ignore the ontological intricacies of truth
and truth conditions regarding 'what is the case'. That is, I shall not further
analyse the question whether truth or 'facts' may exist independently of
human perception and conceptual understanding. Nor whether physical facts
do exist whether or not we know them, whereas social facts are always
110
Cognition
constructed, and hence cognitively and socially relative. In the commonsense world, things and facts are simply assumed to exist whether we know
them or not. Linguistic or cognitive relativism or constructionism are not
characteristic of lay epistemics. 3
We have seen that knowledge presupposes truth criteria, that is, grounds
for justification, whether they are commonsense criteria in everyday life and
hence a basis for the cultural common ground, or scientific ones in the
specific group beliefs of scholarship. We have also seen that these criteria
are historically, socially and culturally variable — what in one period, group
or culture is accepted as reliable evidence of true knowledge, may be
rejected as unacceptable in others. In other words, at the meta-level of a
theory or philosophy of knowledge, as well as in a social and cognitive
approach, knowledge is by definition relative, given the changing nature of
knowledge criteria.
In the practical, everyday world of each period, group, society or culture,
such relativism would be disastrous. Whether 'objectively' valid in some
cases or not, people need to be able to say that some things are true and
others are false, and that there is knowledge on the one hand, and (mere)
beliefs on the other. That is, they take the existence of most objects and the
truth of many facts of their everyday cultural and natural worlds for granted,
and will allow variable types of doubt or ignorance about other things. They
therefore distinguish between knowledge and beliefs, and between objectivity and subjectivity, where subjectivity is defined in tercos of personal or
group beliefs that are unfounded according to us (our group) or according to
the commonsense truth criteria of the shared culture. Whether epistemologically or sociologically naive or not, such distinctions work 'for all
practical purposes', both for lay people, as well as for the 'professionals of
truth', such as journalists, lawyers and scholars.
Ideological relativism?
Does this (simplified) account of knowledge also allow us to decide about
the nature of the relations between ideology and knowledge? This first of all
depends on our basic theory of ideology. If ideology is the axiomatic basis
of the mental representations shared by social groups, and if ideologies vary
as a function of the various interests (membership, activities, values,
position, resources) of each group, the ideological relativism thesis implies
that what group members know is a function of their ideology.
Obviously, in its strong form this thesis cannot be defended. There is no
doubt that most of the knowledge of most groups is shared by other groups.
Or rather, most knowledge is generally, and socio-culturally defined and —
except for some realms of professional or expert knowledge — not in tercos of
specific groups. Indeed, all intergroup communication and interaction, and
even ideological conflict, presupposes a vast domain of shared knowledge.
Moreover, most of this knowledge is undisputed and taken for granted, as
Knowledge and truth
111
explained aboye. Thus, most people in contemporary Western culture know
what trees, tables, cars, computers and myriad other things are, and
presuppose such tremendous amount of knowledge in their everyday discourse. As we have argued in Chapter 3, most knowledge that people of
different groups have is pan of the cultural common ground and hence
undisputed and taken for granted. Thus, the first conclusion is that, given a
group-based definition of ideology, the strong forro of the thesis of ideological relativism (namely, that all knowledge is ideological) cannot be defended
without changing both the commonsense and the theoretical meanings of the
concepts of 'knowledge' and 'ideology'.
But what about the weak version of the thesis? Is specific group
knowledge ideologically based? The sociologically and politically informed
answer to this question would undoubtedly be affirmative, possibly referring
to a long history of ideologically based 'scientific facts' (for instance about
poor people, wornen, blacks or gays) that obviously were in the interest of
some group, namely, the white male middle class and its scholars. Many
other examples can be mentioned in which what is defined or presented as
knowledge are in fact false beliefs, half-truths or one-sided true beliefs that
favour specific groups, and that are directed against others.
Note that this argument not only applies to false or incomplete beliefs, but
also to true beliefs. Nothing, indeed, can be as persuasive as the social facts
being marshalled by the civil rights movement or the womeñ s movement
when it comes to criticizing discrimination and claiming their rights, as both
critical scholarly research as well as fundamental litigation have shown. That
is, minorities or dissident groups will focus on, and highlight their own
truths, and such knowledge could thus, at least in one sense, also be called
ideological.
If these arguments are correct, we must conclude that the weak version of
the ideological relativism thesis is correct: some knowledge in society is a
function of the ideological position or power of groups. This is especially
the case when such knowledge pertains to the social position of the group
itself, or if it is related to the social issues that define the ideological
opinions of the group. Thus, depending on oné s view about smoking,
different beliefs about smoking may be focused on, emphasized, concealed
or denied. Many examples may be given from public debates about smoking,
as well as about immigration, abortion or nuclear energy. Some of these
beliefs may even be true (according to culturally accepted truth criteria) and
hence qualify as common knowledge, but even then they may still be called
partisan in the context of the other beliefs and attitudes of a group. Their
'facts' may thus not be Ours.
Knowledge or opinion?
One possible objection against Chis conclusion is that group-dependent
knowledge is not knowledge at all, but opinion, so that the argument about
ideological knowledge would be pointless, if not a contradiction. 4 This
112
Cognition
argument may be supported by commonsense uses of the concept of
'knowledge'. Groups that are in conflict and participate in an ideological
debate would not readily grant that what the others believe constitutes
knowledge, but will claim that such beliefs are mere opinions. Por example,
research on patterns of discrimination and ethnic beliefs may confirm the
everyday experiences of minorities in the Netherlands, namely, that racism
is endemic in Dutch society. However, such a conclusion is challenged by
most Dutch (including most social scientists) as being merely an opinion,
and in fact hardly more than a typical anti-racist accusation. That is, facts
may be denied if seen to support the ideological position of the others, even
if such facts are the result of research that has been carried out according to
generally accepted scientific methods which in other research would never
be challenged.
Note incidentally that the concept of 'opinion' used in such accusations
has a broader sense than used in this book, where it only means an
'evaluative belief . In everyday language use 'opinion' is sometimes also
used as referring to 'factual' beliefs that (others think) are false. In the rest of
this chapter, I shall often use the broader, commonsense notion of 'opinion',
in order to have a word that denotes all beliefs that are not true and hence
part of knowledge but are evaluative beliefs or false factual beliefs.
We now are facing a dilemmálf at least some knowledge is ideological,
it will in many everyday situations be challenged as not being knowledge at
but merely opinion. Such judgements presuppose the general definition
of knowledge, that is, that beliefs are only accepted as true if we (also)
accept them as true. In this case, 'we' may be simply (most of) the other
members of a culture, society or group, or a scholar or other outsider judging
the beliefs of such a culture, society or group. In other words, if factual
beliefs are defined as opinions as soon as they are understood to be
ideological (at least by the others), then we are back to square A, that is, that
ideologies typically monitor evaluative beliefs only, and not knowledge. In
fact, we would then only have general cultural knowledge, and not specific
group knowledge. Following this argument, we would again have to
conclude that knowledge is not ideological, simply because the cultural
meaning of knowledge presupposes non-partisan belief: as soon as (even
true) knowledge is socially expressed by an ideological group, it will be
degraded to (mere) belief by the others.
But also this conclusion is problematical. Indeed, each side in an
ideological debate may firmly believe and even be able to prove that their
beliefs are true. If not, I would have to recognize that my own books on
racism feature mere opinions and not knowledge that results from careful,
empirical and theoretical research. Indeed, I would further claim that given
such scientific criteria and my results, I 'know some 'facts' about racism in
the Netherlands, whereas those who simply deny such 'facts' (according to
me) are expressing merely an opinion that is obviously based on nationalist,
ethnocentric or racist ideologies, and not on reliable experience or scientific
research.
Knowledge and truth
113
The problem of this (very realistic) last example is that most people in
society may finnly believe something, while a few others believe the
opposite. Theories of knowledge in tercos of consensus, common serse or
shared cultural beliefs, would in that case be in trouble — critical, dissident
knowledge of a few would then be defined as an opinion by a (vast)
majority. Indeed, many people have burnt on the stake for that reason, and
the problems of Galilei with the Catholic Church have only recently been
resolved, after more than three centuries. Contemporary social movements
and action groups have their own stories about the difficulty of getting their
beliefs accepted as knowledge, and not rejected as mere ideological opinions.
There is another aspect involved in the ideological struggle about knowledge and truth, namely, meaning. That is, different social groups of course
share a vast amount of socio-cultural knowledge, as well as many truth
criteria. This allows members of different groups to understand each other,
to argue and sometimes even to persuade each other. However, given their
different interests, some concepts may be defined in a different way in
different groups. Thus, in the ideological debate about racism in the
Netherlands, it may, well be accepted (given the role of — some — social
scholarship in such debates) that more than sixty per cent of Dutch
employers state that they prefer white men over women and minorities. This
statistical 'fact' may be granted (and thus statistics admitted as a truth
criterion), though seldom highlighted, in a debate with anti-racists. But the
difference of opinion begins where one group considers this fact as proof of
racism, whereas the other group simply does not want to call this fact a form
of racism at all, but at most a forro of prejudice, misguided beliefs or
resentment. Indeed, the other group may define 'racism' only in terms of
ideologies of racial superiority, and as characteristic of the extreme Right.
In other words, 'racism' never applies to 'people like us', so that any
evidence of racism that might be applicable to 'our people' is automatically
disqualified as being ideologically biased, and an unjust accusation. In other
words, it is not the knowledge or its basis here that is rejected as an opinion,
but rather the meaning and application of a concept. And since there is no
'objective proof for the correct use of one specific meaning of the words
that deal with social structures and relations, any use that may not be in our
own interests may be rejected as incorrect or biased, that is, as an expression
of an opinion, so that also its truth criteria do not apply in the same way.
Indeed, many of these tercos are generally seen as involving value judgements anyway, and not as descriptors of objective facts or properties, as is
also true for words like discrimination, dernocracy, conservative, progressive, dangerous, healthy, and so on.
We might therefore further specify that any belief, including factual
beliefs, that implies a value judgement, thereby may become an evaluative
belief or opinion for others. Thus, the concept of racism may truthfully
describe the ethnic situation in the Netherlands. But for both racists and antiracists alike, the terco has a negative implication, so that its use tends to be
114
Cognition
seen as a value judgement. Similarly, the statement that some country is not
democratic may well be intended and used as a factual statement, but given
the fact that it may imply a value judgement, it may also be interpreted as an
opinion, and hence as an accusation. Even obvious descriptive tercos, such as
woman or child, may thus in some contexts be intended or understood as
implying a positive or negative value judgement, and hence as the expression of an opinion instead of a factual belief.
This analysis shows something else, namely, the ideological basis of the
core of much (social) knowledge, such as the very concepts that define such
knowledge. Indeed, if 'racism' has the broad conceptual meaning it has for
anti-racists, others may think exclusively of aggressive, extremist, rightwing racism, or only about explicit racist ideologies, when using this terco.
That is, as soon as part of a concept, when applied to our people, is seen to
be inconsistent with our interests, people also adapt their knowledge and the
language used on the basis of it. Similarly, most feminists will probably tend
to define 'sexual harassment' in broad tercos, whereas many men (and some
women) may find this exaggerated and would accept only overt, blatant and
very aggressive forros of sexual violence in such tercos. In other words, each
group may also have its own concepts and language use, and these would be
ideological as soon as dimensions are added to, or deleted from, the concept
so as to apcommodate the interests of the group.
In sum, one way in which ideologies control knowledge is the way they
monitor conceptual structures and hence word meanings. The question then
becomes: who should define such concepts and meanings? This question
brings us to the relations between knowledge and power.
Knowledge and power
An analysis of the role of ideology in the study of knowledge not only
involves an abstract epistemology or cognitive science, but also many social
dimensions that have to do with the establishment of truth, truth criteria and
what counts as knowledge in society. 5 Power is one of these dimensions. Let
us therefore examine whether such a perspective may resolve the dilemma
between the thesis that says that at least some knowledge is ideologically
based and the thesis that claims that all ideologically based knowledge
should be called opinion, so that knowledge by definition is nonideological.
There are several ways to tackle this issue. The first is to change the
definition of knowledge. Instead of saying that knowledge is justified true
belief , we may say that knowledge for a given culture or society can never
be more than justified belief , whether or not it is objectively true, or
whether or not knowledgeable others now or later think it is true or false.
That is, the combined pragmatic-semantic definition is thus reduced to a
pure pragmatic one, which in fact claims that knowledge is based on the
power of the consensus, that is, on the kind of truth criteria accepted within
Knowledge and truth
115
the epistemic community Thus, for the Dutch community, the dominant
consensus is that the Netherlands is not a racist country, and that those who
claim otherwise are not expressing knowledge, but an ideological or
otherwise misguided opinion.
This solution is also in fine with discourse analytical, microsociological
and ethnographic thinking, which emphasizes the role of knowledge as being
generally presupposed and taken for granted within a society or culture. It
also is consistent with a historical and political approach, which would claim
that what counts as knowledge in any period or community is determined by
who has the definitional or other truth-determining power in society, such as
public opinion, the church, the media or science. This argument will also
correctly predict that if specific minorities, dissidents or individuals express
beliefs they hold (and even prove) to be true, these wilr either not be
believed, or their knowledge will be disqualified as mere opinion, or they
will be prevented from expressing their beliefs in the first place. Of course,
it may well be that such 'deviants' later may prove to have been right, and
hence (from an outside point of view) to have expressed knowledge, but that
does not mean that they were right for the epistemic community.
Given this power over the definition of truth and knowledge, one may
claim that such a consensus is itself ideological while being in the interest of
the community as a whole. But such a position would be inconsistent with
the specific definition defended here — ideologies are defined for groups and
presuppose different (and often conflicting) group interests within the same
community. Of course, if we were to see a whole community (culture,
society) as such a group, this would constitute the boundary case of what I
define as an ideological group (see Part II), and indeed the ultimate form of
ideology, namely, that of the consensus, and a culmination of hegemony if
such a consensus could be established by the elites. The interests defended in
that case are indeed those of the community as a whole, and defended
against any deviant individual or subgroup. Although such a position may be
defended, it would in fact collapse the notion of ideology with that of
societal norms or culture, and it would mean that we are unable to use it in
a more specific inter-group sense.
So, if we maintain the definition of ideology in terms of interests of
different groups within a community, the next question is whether we allow
the definition of knowledge to be group-dependent as well. That is, not only
within the community as a whole, but also within its various ideological
groups, knowledge would then be defined as justified belief, whether or not
it is true, or whether or not other groups or 'independent' truth instances
would qualify such beliefs as opinions. Again, such a position would
correctly predict the use of the notion of knowledge within groups, as long
as it can be justified with the truth criteria accepted within that group.
Obviously, much of the general socio-cultural knowledge holding within
the group may be shared with other groups, and the same may be true for
most truth criteria. But especially the beliefs and truth criteria that are related
to the interests of the group, or the special issues that are relevant for the
116
Cognition
group, might well be specific, and hence ideologically based — again,
whether or not they are 'objectively' true or false. Thus, the specific
knowledge of women about sexual harassment, of feminists about gender
inequality, of anti-racists about racism, or of ecologists about pollution,
constitute relevant examples. 6 Again, in these cases such knowledge may
well be objectively true (given an independent truth instance), but it may be
rejected as opinions, lies, or fantasies by those who oppose such groups.
Conversely, their opponents may firmly believe — and never see as mere
opinion — things about gender, immigrants or pollution that are objectively
false. That is, what counts is what the group members believe and what,
within their own system of verification, they believe to be justified — whether
or not such truth criteria are themselves biased. A typical example is
knowledge about ethnic relations. Dominant majority group members may
feel that any knowledge and epistemic criteria in this case as defined by
minority groups will be biased. This is, for instance, the case for those
(many) white journalists who do not take minority sources and their
statements seriously. In other words, the basis of credibility judgements
themselves may be partisan, and hence ideological. This also explains why
specific group knowledge of one group will often be rejected as mere
opinions by opposing groups. Indeed, very often the very ideological
conflict itself may not only be about socio-economic conditions or resources,
but about truth criterio themselves.
Since many ideologies are constituted by fundamental opinions about Us
and Them, we must assume that not only the basis of attitudes are evaluative
beliefs, but also those of specific group knowledge. That is, although within
the ideological group, knowledge is still distinct from opinion, the knowledge criteria themselves are self-serving and value-oriented. For instance,
such criteria may involve (value) judgements about who is a reliable source,
what is relevant information, what perceptions can be trusted, or what data
can be depended on. Thus, Christians may admit God as one of the instances
of truth, and anti-racists the everyday experiences of minorities in a racist
society.
Concluding remark
Concluding this succinct discussion of the role of knowledge and truth in a
theory of ideology, we thus find again that ideologies in general monitor
group attitudes — that is, evaluative beliefs — but that also specific factual
beliefs may be defined as knowledge within the group. That is, ideologies
essentially control group specific judgements about what is good and bad,
and also about what is trae or false for us.' This may also include parts of the
meanings of specific concepts (such as 'racism'). This does not mean that,
from an independent point of view, all group knowledge is ideological, since
each group obviously shares knowledge with other groups. Nor does it mean
that all truth criterio are ideological, since each group must be able to argue
Knowledge and truth
117
in such a way (using general truth criteria) that others can be persuaded of
their position.
Ideological knowledge control, however, does consist in selecting concepts and truth criteria that may be specific to a group, and may involve
attributing special credibility to specific truth instances, such as God,
Science, the Party or the Union. This also mean that again within the group
itself such partisan knowledge is not found to be 'ideological' (and hence
misguided) at all, but knowledge like any other kind of knowledge. But
since group values, principies and other basic beliefs are involved that
reflect the interest of the group, our (outside) description of course would
generally take such knowledge and its truth criteria to be ideological, as
defined.
12
Identity
What is identity?
Ideologies consist of a fundamental schema of which the first category
defines the membership criteria of a group. Together with the contents of the
other categories, such criterio define the social identity of a group. This
means that whenever a group has developed an ideology, such an ideology at
the same time also defines the basis for the group's identity. The question is
what this implies exactly. Does it mean that group members can only be
considered group members, and hence partake in a group's ideology when
they actually identify themselves as group members? What exactly 'is' such
an identity, and the process of 'identification' in the first place?
Again, my approach to such questions is socio-cognitive: identity is both
a personal and a social construct, that is, a mental representation. I briefly
discuss this element in the theory of ideology precisely because it may be
situated at the boundaries of a theory of social identity, a theory of social
cognition and a sociological theory of group membership. I
In their representation of self, people construct themselves as being a
member of several categories and groups (women, ethnic minorities, US
citizens, journalists, environmentalists, etc.). This self-representation (or
self-schema) is located in episodic (personal) memory. It is a gradually
constructed abstraction from personal experiences (models) of events. 2
Since such models usually feature representations of social interaction, as
well as interpretations of discourse, both experiences and their inferred selfrepresentations are at the same time socially (and jointly) constructed. Part
of our self-representation is inferred from the ways others (other group
members, members of other groups) see, define and treat us. When experiences are shared with others, abstracted personal experiences, and hence
self, may pardy merge with the self-representation of the group. A feminist
may thus feel herself to be a feminist in more or less the same way as other
feminists do, and, in that respect, self of an individual feminist may be pardy
constructed with the elements of the socially shared self-schema of feminists
as a group. The more the feminist construction of self corresponda to the
socially communicated and shared group schema, the more an individual
woman will 'identify with feminism.
This does not mean, of course, that such a weak or strong group
identification needs to be dominant in specific events and situations. A
feminist joumalist, when gathering or writing news stories, may well
Identity
119
primarily identify as a journalist (and hence adopt ideologically based
joumalistic attitudes, including the opinions and models derived from those)
and only secondarily as a feminist, and the converse will be true if that same
woman participates in feminist action.
In other words, group identities may be more or less abstract and contextfree, in the same way as 11 social representations are. Similarly, social
members may share in several social identities that are more or less stable
across personal contexts, and thus defining a personal self, but in concrete
situations some of these identities may become more salient than others.
Thus, in each situation, the salience, hierarchy or relevance of group
identification will monitor the actual social practices (e.g. the action priorities or 'motivatioñ) of social actors. Unless we admit a theoretically
doubtful notion such as 'situational identity', thus, we should distinguish
between relatively context-free personal identity (which may be a composition of various social identities) or personal self, on the one hand, and the
actual, situated practices of social actors that may be seen as manifestations
of (some aspects of) personal identity.
People may'objectively' be members of groups (and be seen by others as
group members) and still not identify with their groups. Such well-known
forms of dissociation, which most dramatically may occur for inherentidentity groups (young, old, men, women, whites, blacks, etc.), but also for
professional groups, probably implies that such 'members' do not share the
ideology of the group either. Indeed, they may, for a number of reasons,
rather identify with opposed groups and their goals and values. Derogative
words such as 'traitor', 'renegadé , 'dissidene, 'linde Tom' and so forth
show what kind of reactions and sanctions group members may face when
denying or leaving their own group. It also explains why anti-racists are
sometimes considered to be more of a problem in white society than racists
— they share the ideology of the others that'our' society is racist, and thus
threaten the positive self-definition of 'us' as the dominant group (see
Chapter 28). Treason is, either literally or at least symbolically, a capital
offence for many groups, as is the case for sedition, defection, and becoming
an'infldel'. Conversely, strong identification and co-operation will usually
be positively valued in tercos of solidarity, allegiance and fidelity. AH this
applies not only to social practices, but also to ideologies and the forms of
'mental' solidarity with groups represented in self-representations that may
be assumed to be at the basis of such social practices.
Personal and group identity
These arguments suggest first that we need to distinguish between social or
group identity and personal identity. The latter takes the two forms informally described aboye: (1) a mental representation as (personal) self as a
unique human being with its own, personal experiences and biography — as
120
Cognition
represented in accumulated mental models, and the abstract self-concept
derived from it, often in interaction with others; and (2) a mental representation of (social) self as a collection of group memberships, and the identification processes that are related to such membership representations. These
identification processes may be assumed to depend on a comparison between
personal and social self. If the membership criteria, activities, goals, norms,
values, position or resources of the group are in fine with (at least consistent
with) those of the personal self construct, identification may be more or less
strong. If not, a process of dissociation may take place, including association
with other groups.
For a theory of ideology, this of course has implications for the ways
individuals identify with group ideologies and attitudes. When membership
is largely ideological (as for political parties, Churches, etc.), such ideological dissent usually implies leaving the group altogether when one's dissenting opinions are inconsistent with those of the group as a whole. For
professional ideologies this is much more difficult because they are closely
related to goals and interests of everyday professional practices. It is difficult
to 'be' a professor and at the same time not 'feef like one, and if professional ideologies represent the aims, values, norms and social resources
of the professional group members, ideological dissociation is seldom in
one's personal interest. Of course, there may be other considerations, other
ideologies and values, which may be accepted as more valid, despite one's
group membership. Thus, occasionally professors may espouse student
ideologies.
Ideology as group identity?
All these processes account for personal variation and the complexity of the
manifestations of group ideologies in everyday life. However, it should be
recalled again that ideologies are essentially shared and hence need to
be defined at group level. The same is true for the social or collective
'identity' of the group as a group. Usually, identity is taken in an individualistic fashion, in terms of representations and identification processes of
group members. However, in the same way as groups may be said to
share knowledge, attitudes and an ideology, we may assume that they share
a social representation that defines their identity or 'social self as a
group. 3
My attempt to bring some clarity to the multitude of notions related to the
field of ideology suggests that, at least at the cognitive level of description,
social (group) identity probably collapses with a group self-schema. And
since I have taken such a schema as the most likely candidate for the format
of a group ideology, we need to conclude that group identity collapses with
group ideology. 4 Given the way I have analysed social ideologies, this is not
entirely improbable, sine the relevant categories precisely define what
Identifies' the group, especially also in relation to other groups. That is, the
Identity
121
ideological group self-schema should represent precisely those fundamental
beliefs that are generally shared (acquired, used, reproduced) at the
group level, and answer such fundamental questions such as 'Who are we?',
Where do we come from?', 'Who belongs to us?','What do we (usually)
do, and why?', 'What are our goals and values?', and so on. 5 The theoretical (general, ideological) answers to such questions are therefore continuously taught and repeated in social encounters, symbolic interaction, and
other group activities. It is this that is inculcated, sometimes explicitly
(in didactic situations or in times of crisis), and often implicitly; in the many
group-relevant social practices of the group, its institutions and its
members.
On the other hand, there are a number of arguments that plead against
equating group identity with ideology. Thus, if the cognitive dimension of
group identity is defined in terms of the specific social representations shared
by the group, the notion of group identity is more inclusive than that of
ideology. After all, ideology was more strictly defined as the 'axiomatic'
basis of the shared social representations of a group. That means that
ideologies form at most the basis of group identity, that is, the fundamental
propositions that pertain to more or less stable evaluations about 'our'
group's membership criteria, activities, goals, norms and values, social
resources and especially our position in society and the relations with special
other groups.
Just like personal identity, social identities may change. Whereas some
basic (ideological) principles may remain more or less identical over a
relatively long period of time, the more specific social representations, such
as attitudes, may adapt strategically to social and political change. Thus,
although the peace movement might of course keep its basic pacifist
ideological principies, specific attitudes about different forms of disarmament, the deployment of nuclear arms, and other issues may depend more
directly on the political situation, including the changing attitudes of
opponents, or the realization of oné s major goals. 6
Such changes of group attitudes more generally pose the question about
the nature of social identity. If social identity is defined in terms of shared
social representations, and if these may continually change, also the very
notion of identity should be more a dynamic than a static notion. But if
social group identity is in turn a crucial defining property of social
movements and other groups in the first place, then the very notions of
movement and group need to be much more dynamic. As we shall see in
more detall in our discussion about groups, in Chapter 15, this would mean
that a group is not merely a more or less stable collectivity of people, but
also or rather defined in terms of a permanently changing set of cognitions
and their concomitant practices. Identity then becomes a process in which
such a collectivity is engaged, rather than a property. For that reason the
term identification would probably be more satisfactory than the more static
term 'identity'. Just like persons, groups may thus be permanently engaged
122
Cognition
in a 'search' for their identity, as a function of social structures as well as
changes.
Social identity as collective feeling'?
Social identity is often associated also with more affective or emotional
dimensions. Although these concepts open the well-known Pandora-box of
emotion theory, and raise the old question whether emotions are (also)
cognitive and not (merely) physiologically based, we should not run away
from such theoretical problems. The point is that if emotions necessarily
have a physiological basis, they need to be strictly personal, since groups
obviously have no bodies. But, similarly, groups have no minds, and we still
speak about socially shared mental representations. Thus, what does it mean
that group members may share 'emotions' as distinct from sharing (strong)
evaluative beliefs?
If emotions are (also) defined in tercos of bodily arousal of some kind,
then a 'shared' emotion would imply that group members would be
constantly aroused. Thus, if feminists are 'angry' about male chauvinism,
does that mean that afi individual women who identify themselves as
feminists constantly 'feel' angry. Of course not. However, individual feminists may be (more) likely to become angry at specific moments of expressions of male chauvinism. But that is not the same as saying that feminists
as a group (permanently) 'share anger', in the strict sense of an emotion.
Rather, I would suggest, such an expression does not denote emotion at
but strong negative beliefs. Indeed, while holding such strong negative
beliefs, some feminists may never actually feel angry about social inequality
of the group, although again they may become angry about personal
experiences of such inequality. The same is true, more generally, for feelings
of social identification. One may 'feel' strongly about oné s membership of
a group, but again such a 'feeling', I propose, is a set of evaluative social
representations (e.g. attitudes about equal pay, abortion, etc. for feminists),
rather than an ongoing emotion shared by al 1 or most members of the
group.
In other words, the frequently observed emotional attachment of members
to a social group may not be, as such, an alternative to the cognitive
definition of group membership given aboye. This does not mean that
individual group members may not tend to be (more) emotional in their
personal (but group-related experiences). However, it does mean that such
emotions, as such, cannot be actually 'shared'. They may be known,
respected, talked about, and in that way they are 'shared'. But there is no
such thing, it appears, as a 'collective emotion' of a more or less permanent
nature. This again does not mean that, at a specific moment, a collectivity of
people may not have more or less the same emotion, for instance when
demonstrators are angry during a demonstration. But that is not the same
thing as a shared, collective feeling of a group, a feeling that exists also
beyond such specific 'emotional' moments.
Identity
123
Other mean of social identification
However, unlike ideologies, social identities need not, as such, be limited to
the cognitive realm. Group identity may also be defined, at least partly, in
terms the characteristic social practices of group members, including collective action. Indeed, members of a social movement might identify as
much with the 'ideas' shared by the group, as with such typical group
activities as demonstrations, strikes, meetings or rituals. Initiation rituals
may indeed be a major criterion of group membership and hence of (feelings
of) identification. The same is true for group-identifying symbols, such as
uniforms, flags, badges and many others. Again, theoretically one might see
both social practices and these symbols as expressions or manifestations of a
more abstract, 'underlying' group identity, as we have done for ideologies.
However, the personal and social processes of identification and sharing
are not limited to such abstract, cognitive representations. In order to avoid
reduction of group identity to specific actions or ad hoc symbols, we might
require that the practices and symbols involved should also have a more
permanent, general or routine nature. Indeed, uniforms and flags typically
have a more permanent character. And group identification with, for
example, demonstrations seems more likely when such demonstrations are
more or less characteristic of the group, and not when they occur just once or
twice. An apparent exception to this rule are prominent historical events that
contribute to group identity, such as the Russian Revolution for communists
or the March to Washington for the civil rights movement. Thus, also many
nationalist movements tend to search for famous historical events, historical
figures, monuments, places as symbols of group identity. Precisely given
their historical nature, they have become preserved as parts of collective
memory, and hence again qualify as a criterion for identification. 7
These well-known examples suggest again that group identity does not
seem to be limited to shared mental representations, but involves a complex
array of typical or routine practices, collective action, dress, objects, settings,
buildings (like churches), monuments, prominent historical events, heroes
and heroines and other symbols. At the same time, a more cognitive
approach would in that case emphasize that it is not so much the symbolic
actions and objects themselves that are criteria for identification, as rather
their collective social construction, that is, again some form of shared
representation. It is not the material form or substance of the cross that
defines Christian identity, but the complex, interpreted 'story' that the cross
symbolizes and that Christians share. In other words, where group activities
may suggest that identification is based on collective action or relevant
objects, further analysis suggests that precisely the 'symbolic' nature of such
phenomena requires at least also a cognitive analysis in terms of the socially
shared interpretations assigned to such collective actions and symbolic
objects.
This conclusion does not imply that all criteria of social identification are
merely' mental. Apart from the actual social discourses and other practices
124
Cognition
in which group members may engage, as well as objects, settings and other
properties of collective events, also various kinds of social structure and
organization may be involved in identification. Thus group identity may also
depend on official membership activities, such as asking and paying fees,
electing officials and leaders, institutionalizing a movement, and so on. We
shall deal with these (more) social dimensions of groups in Chapter 15, but
it should be emphasized here that the cognitive dimensions of 'feeling' that
one is a group member, as well as the shared processes of group identification, may also be related to social practices, organization and institutionalization. Indeed, you may really feel you are a member as soon as you get
your membership card. Not surprisingly, one lexical manifestation of the
relation between group identity, ideology and institutionalized membership
is obvious in such expressions as whether people are 'card-carrying'
members of a movement or not.
There is one major problem when we extend the notion of social identity
towards the vast world of social practices, symbols and organization — it
would make the notion of identity as comprehensive and vague as that of
culture. In that case, social identity might even collapse with that of groupculture. That is, in the same ways as members of a larger national or ethnic
culture would identify with their culture, a similar process would exist for
social groups. Since the notion of social identity has no fixed meaning, we
might simply adopt such a broad definition, but somehow we seem to be
overextending the notion in this way. We would probably also hesitate to
extend the notion of personal identity to all actions, dress, personal objects,
and so on, of a person, although also here, it would simply depend on
whether one opts for a broader or a more specific view of identity.
Characteristic actions, ways of speaking or dress could of course be taken to
define a person' s identity.
The conclusion from this discussion should be that, as is the case for
ideology, social identity is a very fuzzy notion, and it simply depends on the
theorist whether a strict or broad perspective is taken. I tend to opt for the
stricter and more precise definition. That is, in the same way as we
distinguish between ideologies as such, on the one hand, and the many
manifestations of ideology in discourse, interaction and ideological symbols,
on the other, we máy thus restrict social identity as such to a shared core of
social self-definition, that is, to a set of social representations that members
consider typical for their group. The social practices, symbols, settings or
forms of organization that are typical for a group and with which members
identify, would in that case be the contextually variable manifestations of
social identity. In line with the subjective nature of 'feelings of belonging' or
'commitment' with respect to a group, such a socio-cognitive definition
would also explain that it is not so much á social practice, symbol, setting or
organization itself that is part of a social identity, but rather their meaning or
interpretation for the group.
This definition of social identity as a socially shared mental construct also
allows for individual variations of interpretation, historical changes in the
IH
Identity
125
meaning of the 'externar manifestations of social identity, as well as for
processes of socialization of members at the individual level and group
formation at the social level. Indeed, different groups may be associated with
the same type of social activities, objects, symbols, settings or organization
forms, but attach totafiy different meanings (social representations) to them,
and thus construct a different kind of social identity. In that sense, social
identity is as intersubjective as personal identity is a subjective construction,
although both constructs are obviously also a function of social interaction
and negotiation, and of the attribution of identity by other people and other
groups, respectively.
Finally, this socio-cognitive approach to the analysis of social identity
also allows a systematic relationship with the role of discourse in the
construction of social identity. 9 An important part of the formation and
reproduction of social groups may indeed have a discursive nature. Social
groups in general, and social movements in particular, are constituted by
various forms of intragroup discourse, such as meetings, teaching, calls for
solidarity, and other discourses that define the ongoing activities, the
reproduction, and the unity of the group. On the other hand, social group
identity is especially also construed by intergroup discourse in which groups
and their members engage for reasons of self-presentation, self-defence,
legitimation, persuasion, recruiting, and so on. Although it was suggested
aboye that I prefer to distinguish between social identity itself and the social
practices, including discourse, based on such an identity, it is obvious that
group discourse is a rich source for the analysis of 'underlying' social
identities.
13
Social Cognition
The relevance of social cognition
Having completed the first part of the theoretical framework for the study of
ideology, let me take stock of the relevance of the cognitive component in
such a theory, and then discuss some open problems and further prospects.
The main arguments that have given rise to a cognitive component have
been the following•.
1 Whatever else ideologies are, or whatever social conditions and functions they have, they are first of all systems of beliefs. The nature of
these belief systems, as well as their relations with other mental objects
and processes, (also) need to be studied in a cognitive framework.
2 Ignoring such cognitive dimensions of ideologies, and merely analysing
them in tercos of social practices, social formations, or social structures,
provides incomplete insight into ideologies, and constitutes an improper
reduction of complex social phenomena, and hence an inadequate
theory.
3 Ideologies are socially acquired, shared, used and changed by group
members, and hence are a special type of socially shared mental
representations.
4 Ideologies are reproduced through their everyday uses by social members in the accomplishment of social practices in general, and 'of
discourse in particular. Such uses not only have social foundations but
also cognitive ones, such as the personal experiences, knowledge and
opinions of social members. In order to relate the social dimension of
ideologies with their personal uses, only a cognitive theory is able to
provide the necessary interface.
Beyond 'belief systems'
We have seen that ideologies are not just any set or system of ideas or
beliefs, because iri that case they would simply coincide with cognition in
general. Nor should ideologies be reduced to the social knowledge, attitudes
or 'worldviews' individual people have. Rather, ideologies form the 'axiomatic' basis of the shared social representations of a group and its members.
That is, they are both mental and social phenomena.
It is this integrated socio-cognitive aspect of ideologies that is the core of
the theory presented in this book. Though traditionally associated with
Social cognition
127
mental notions such as 'ideas', 'beliefs', 'consciousness', 'common sense'
and related notions, these mental dimensions of ideologies have rarely been
analysed in any detail in most philosophical and sociological studies of
ideology. Similarly, the psychological work on ideologies has paid attention
to'belief systems', but these were hardly analysed as such, but rather used
as an independent or dependent variable in the explanation of social or
political'behaviour'. The same is true for socio-historical studies about the
ideas or ideologies of specific groups or periods, although such studies
obviously provide an interesting empirical basis for further analysis of
underlying ideological systems.
In this framework, a cognitive analysis first establishes the nature of the
theoretical components of ideologies, namely, specific beliefs. Theoretically,
such beliefs are traditionally represented as propositions, although other
formats might be envisaged as long as they are able to account for the
general and abstract nature of ideologies.
That ideologies are not merely the 'beliefs' of a group may also be
concluded from the following list of different kinds of beliefs people may
have:
1 knowledge (factual beliefs) of individual persons about particulars
(people, objects, events, etc.)
2 knowledge of individual persons about categories or classes of particulars and their properties
3 opinions (evaluative beliefs) of individual persons about particulars
(people, objects, events, etc.)
4 opinions of individual persons about categories or classes of particulars
and their properties
5 knowledge of social groups about particulars (people, objects, events,
etc.)
6 knowledge of social groups about categories or classes of particulars
and their properties
7 opinions of social groups about particulars (people, objects, events,
etc.) .
8 opinions of social groups about categories or classes of particulars and
their properties
9 social beliefs of a whole culture (cultural common ground)
10 norms, values and truth criteria as the basis of the cultural common
ground.
Ideologies as social representations
Against the background of a critique of traditional approaches to ideology, it
was first decided to limit ideologies to socially shared representations of a
general and abstract kind. That is, ideologies are of the same family as
socially shared knowledge and social attitudes. Ideologies are not individual
128
Cognition
and not represented like specific, episodic memories, or as personal opinions. This is also why the comparison between ideology and language (or
grammar) is so instructive. Both are abstract social systems shared by groups
and used to accomplish everyday social practices, namely acting and
communicating, respectively.
This group-based nature of ideologies and the social beliefs they control
explains how and why social actitudes may be organized as coherently
structured sets of group opinions. Since we typically may disagree about
opinions and different groups may have different or conflicting goals or
interests, it is not surprising that the ideologies that underlie such opinions
are associated with groups.
For other social beliefs, such as knowledge, this appeared to be less
straightforward, simply because knowledge that is associated with a group is
often described as a partisan opinion. For that reason, we distinguished
between general, taken for granted, consensual knowledge of a culture, on
the one hand, and the factual beliefs of a group, on the other hand. These
beliefs group members (with their own truth criteria) may call knowledge,
but others may see them as 'meré beliefs or opinions. It is this specific
group'knowledgé that is controlled by group ideologies. In other words,
knowledge is always by definition relative, that is, described as 'true'
relative to a group or to a whole culture, according to che truth criteria of that
group or culture.
Since ideologies represent the 'axioms' of social group beliefs they are
relatively permanent. Even less than actitudes and group knowledge, and
certainly less than personal beliefs, they do not change overnight. Given
their position in the system, their change would involve the change of a vast
part of the social representations of most members of a social group, and
such a change usually takes a lot of time.
The structure of ideologies
Once ideologies are defined as the foundation of group representations, we
need to examine their intemal structures. What kind of abstract social beliefs
are involved here, and how are friese organized? As everywhere in the
cognitive system, effective processing and uses in social practices requires
organization, for instance by abstract schemata consisting of a number of
categories. That is, if people have to learn, use and eventually change many
ideologies in their lives, at least as many as the groups they belong to, then
it is likely that they acquire and use a special ideological schema to do so.
Since no such schema is available from other domains of cognition, I
provisionally proposed a schema featuring categories that would represent
the essential social dimensions of a group, namely, membership, activities,
goals, values, relations to other groups and resources. This tentative schema
should be used to represent the fundamental opinions group members have
Social cognition
129
about themselves, as well as about their position in society. In other words,
a group self-schema is the core of all ideologies. Thus, racism as an ideology
is first about who We (white people, Europeans, etc.) are, what we look like,
where we come from, what we stand for, what our values and resources are,
that is, what our interests are, and how they relate to those of a specific other
group, namely, non-whites.
The familiar polarized nature of the expression of ideologies, as Us and
Them, reflects the position (or group relations) category of such an underlying structure. This schema also explains the essential group-based, and
self-serving nature of many ideologies, as representing not only the interests
of a group, but also its social position and perspective on any social issue
that is relevant for it. This relevance is again measured in relation to the
fundamental beliefs of each category, such as membership, goals or resources. Any social event or arrangement that may be at odds with these
essential group interests will thus be judged negatively, and such negative
judgements are used as the basis for negative social action, such as
discrimination.
From social representations to personal models
Finally, precisely in order to be able to relate such abstract and fundamental
forms of social cognition to the particularities and realities of situated action
and discourse, another interface is necessary to translate and connect social
opinions with the personal ones of individual social actors. After
although ideologies as such are social and shared, they are actually used and
reproduced by individual group members and in specific social practices.
Therefore, the important notion of a mental model was used to act as this
interface between the social and the personal. Models represent specific
events and actions, and hence account for unique experiences in specific
contexts, but at the same time they embody instantiated ('applied) versions
of social knowledge and opinions as they are derived from knowledge and
attitudes. That is, via the more specific social opinions of attitudes (e.g.
about affirmative action), individual group members may form their own
personal opinions, as represented in models, about concrete instances of
affirmative action, and act upon (speak about) such opinions. Various kinds
of model forro the basis of action, text and talk, and thus provide the
interface that allows ideologies to be expressed and reproduced.
With this, admittedly still sketchy, framework we at least have a coherent
theoretical 'chain' that links social structures, including groups and group
relations (e.g. of domination), via ideologies to other social representations,
and the latter again with models, which finally provide the missing link with
discourse and action. And conversely, we now have the mean to describe
and explain how ideologies — and social relations — may be produced by
discourse and interaction and their cognitive consequences.
130
Cognition
Relevance of the theoretical framework
The framework also allows us a somewhat more explicit discussion of a
number of classical issues in the philosophy of ideology, such as the truthfalsity debate, or whether a critical concept of ideology should be restricted
to ideologies of domination I have provisionally answered both questions
negatively: ideologies are not primarily about what is true or false, but about
how people represent their beliefs about themselves and about the social
world, truthfully or not. The criterion is not truth but relevance (self-serving
social functions, interests). In other words, and somewhat loosely, we may
say that we need a pragmatics of use of ideology rather than a semantics of
truth. The same is true for the restriction to the use of ideologies to
reproduce power abuse and domination.
Obviously, ideologies are often developed and used to sustain and
legitimate domination, and such uses invite critical analysis. But the
interesting and theoretically much more attractive thing to do is to match
domination with resistance, and ideologies with counter-ideologies, for
example sexism with feminism, racism with and-racism. I have argued that
there is no good theoretical reason why the second parts of there pairs
should not also be ideologies. That may require some conceptual adjustment
with respect to the traditional notion of ideology, but it surely is a more
adequate approach, while at the same time not blunting the critical dimensions of the traditional (Marxist, neo-Marxist) approaches.
In sum, adding a powerful cognitive dimension to the philosophical and
sociological tradition, and relating both to a more concrete discourse
analytical approach, allows us to design an analytical framework that one
day may lead to a proper 'theory' of ideology. This will allow us to both
describe and explain in detall how exactly members of specific groups speak,
write and act ideologically. Instead of the more global, macro approaches to
ideologies in tercos of belief systems, hegemony or social formations, this
approach spells out the structures, the everyday uses, the cognitive and
social functions, the acquisition and the changes of ideologies within such a
broader societal context.
Other approaches
Interestingly, precisely also in the social and critical approaches to discourse
and ideology, ideologies have been rife.' Instead of self-critically examining
which theories, concepts and methods are most adequate and effective, the
dominant ideology in the study of ideology says that cognitive science is at
the wrong (scientistic, positivistic) side of the fence. Linguistics, at least
among linguists who have become critical analysts of ideology, is more
acceptable, if only as a useful instrument, or because the social dimensions
of language can be emphasized. But among the philosophers and sociologists, both were irrelevant or suspect, or simply unknown or ignored.
Social cognition
131
Those who are justly opposing the limitations of much contemporary
psychology, have thrown out the cognitive child with the bath water,
although, just like discourse and language, cognition or mental representations can be as 'social' as any concept in the social sciences.
The price that has been paid for this ignorance or ideological exclusion is
that analysis of the way social group members actually go about talking or
acting ideologically has been reduced to an account that fails to link social
structures with cognitive structures and again with discourse structures.
Apart from the misguided accusation of individualism, cognitive approaches
are also rejected because of the thesis that these are mentalist, and hence
opposed to the 'materialism' required in the (Neo)Marxist paradigm, or the
'interactionism' that governs much current work in ethnomethodology or
discursive 'psychology'.
In interactionism, the mind is either seen as a figment of the dualist (mind
versus body) imagination, or as irrelevant because what counts, socially, for
social members, is what is 'observably' being displayed. This neobehaviourist misconception is hardly much more sophisticated than the old
version of behaviourism that has marred psychology and the social sciences
for decades. Yet, at the same time, obviously mental concepts for hardly
observable things like meanings, understanding, rules, and so on, keep
appearing, unanalysed, in such interactionist approaches, as if 'displays' of
meanings or understandings are more observable than these meanings or
understandings themselves. Yes, in common sense they are, and as such are
used as evidence (1 have seen this myself, , 'I have heard this myself ),
simply because the socially shared concepts that govem perception are taken
for granted in commonsense observation. But, if common sense should be
used as evidence, then we should also use the commonsense acceptance of
the obvious presence of meanings, intentions, knowledge and opinions as
properties that people'have in mind'.
And why discourse expressions, action, social practices, social or economic conditions, interest or power are more 'materialist' than meanings and
understanding, has been accepted by fiat instead of by investigation. Any
adequate epistemology will tell us that all these things are both socially and
mentally constructed — actions or discourses cannot be observed any more,
and are no more material, than meanings, knowledge, opinions, values or
ideologies. No interactionist or materialist discourse analyst or sociologist
goes down to the level of physical or biological body movements to describe
social action. Given the concepts and knowledge of our culture, social
actions are themselves conceptual constructs paired with these physical
observables of the movement of body and mouth. Their understanding by
group members is no more immediate than 'underlying' meanings, as
frequent observations of ambiguity or vagueness of discourse or action also
show.
That is, both the social and the cognitive notions are abstract constructs of
everyday understanding, action and mind, as well as of their non-naive
theories. None of them is more or less 'material', 'observable' or otherwise
132
Cognition
more relevant because being socially'clisplayed'. Understanding what
people 'observably' do or say, is also an interpretation of both lay participants and scholars. That such interpretations are being acquired, used,
changed, negotiated in social situations and in social interactions, is obvious
(and a good reason to criticize psychologists who in turra ignore that
dimension), but does not mean that therefore cognition is irrelevant.
On the contrary, all these interpretations and the knowledge and opinions
on which they are based, are themselves both mental and social, depending
on the scope or the level of theory and analysis. Discourse is the most
obvious example in point, while obviously involving mental representations
(meanings, knowledge, abstract structures at many levels), and at the same
time being a form of social, political or cultural action. In sum, social and
cognitive analyses of discourse and ideology that mutually ignore each other
are doomed to produce incomplete, reductive or plainly misguided theories
and analyses.
This conclusion does not imply that we should blindly accept all theories,
methods or philosophies of contemporary cognitive psychology and cognitive science, nor the mainstream orientation in cognitively inspired social
cognition research in social psychology. Overall, this research has been
justly criticized for its fundamental lack of accounting for the social
dimensions of the mirad, its individualism and its mentalist reduction.
Similarly, mainstream social psychology has ignored the fundamental role of
discourse in the construction of the social mirad. On the other hand, from a
theoretical point of view, both social cognition research as well as research
on social representations may be criticized for the simplicity and the
vagueness of their analysis of mental structures and processes. And finally,
virtually all psychology (except the study of political cognition) has ignored
the fundamental role of ideology in the control of social representations and
social interaction.
Open problems
Of course, also the framework presented here is hardly complete. The
schema designed to represent ideological structures is very tentative, and I
am not sure it allows representation of all types of ideologies, especially
those (like environmentalism) which seem to focus more on nature than on
groups, or vast ones, such as communism or religious systems, which have
the whole world as their scope. Also, the schema may be too simplistic.
Complex sets of ideological beliefs may need more structure than that of a
simple schema, although the routine application of ideological principles in
everyday life probably does not allow a very complex structure either.
Next, it was assumed that ideologies organize and monitor more specific
group knowledge and attitudes. But we have only the faintest idea how that
happens (or how, conversely, ideologies are derived from specific social
beliefs). The lack of sophisticated theories for the structures of social
Social cognition
133
representations in general, especially also of attitudes, is another problem.
Much empirical work on concrete expressions of ideologies in discourse
may be necessary to reconstruct such 'underlying social representations.
Many of these problems are related to our fragmentary knowledge about
the organization, the contents and the processing of social beliefs in general.
Some of these traditional problems, such as those that pitch stability and
continuity of attitudes or ideologies against often-observed variation, contradictions and dilemmas, have in my opinion been theoretically resolved with
the introduction of event and context models in episodic memory. These
models also explain the classical cleft between the macro and the micro, the
social and the personal, and provide the interface between ideologies and
social practices. I consider this element in a general theory of ideology as
crucial and as one of the major new ideas this study would like to
propose.
Other problems remain, however. Some of these can be resolved by
empirical work, not only in the laboratory, but especially also by detailed
analyses of manifestations of ideologies in discourse and social practices.
However, it is unlikely that problems of mental structure and organization
can simply be solved by more or better observation. At present there is little
hope that neurological (brain) research will soon deliver the underlying
building blocks that will explain the intemal organization of social representations. This means that we need to satisfy ourselves with the more
abstract, higher-level analysis in terms of cognition.
As is the case for all 'non-observables', theoretical cognitive modelling is
the crucial answer, and will allow us to find more elegant ways to account
for the 'data' (discourse, social action, social organization, social processes,
etc.) at hand. Notions such as 'models', 'scripts', 'schernata' and 'social
representatioñ precisely are the result of such theoretical endeavours. The
same is true for my attempt to devise a more detailed theoretical concept of
ideology as the basic framework of social representations.
Besides the fundamental problems of mental architecture and organization, a socio-cognitive theory of ideology needs to account for actual
acquisition, uses and change. A major role in such a theory is again played
by mental models that serve as the interface between ideologies and other
social representations, on the one hand, and everyday experiences and
practices, and especially discourse, on the other hand. That is, models form
the missing link of a cognitive theory of the actual acquisition, uses,
implementations and changes of ideologies. They explain how social members produce and understand action and discourse and how in turra such
processes are linked with socially shared beliefs, and hence with ideologies.
However, we still have limited insight into the ways contextualized
personal experiences and practices are béing shared, normalized and
accepted at the'aggregaté level of groups. Discourse and (mass) cornmunication again play a fundamental role here, but we should not forget that
accounting for discourse production and understanding is a description of
134
Cognition
what social members do, and not what whole groups do. Sharing beliefs
interactively is one thing, but sharing beliefs throughout a group is another,
hardly less complex phenomenon, especially if we do not want to reduce
such sharing to a mere accumulation of individual learning and interaction.
In order to solve some of the problems mentioned aboye, we therefore
need to take a closer look at the social dirnensions of ideologies, and then
examine how the combined cognitive and social approach can be validated
by detailed discourse analysis.
Part II
SOCIETY
14
Ideology and Society
Relating the cognitive and the social
Whereas the first part of this book has made a strong plea for the
incorporation of a cognitive component in a multidiscipfmary theory of
ideology, no such plea is necessary in this second part for a social approach
to ideology. All traditional approaches agree that ideologies are social, if
only by their multiple social conditions and functions. 1 Even in my cognitive
approach; this social dimension has been emphasized. Ideologies are not
merely sets of beliefs, but socially shared beliefs of groups. These beliefs are
acquired, used and changed in social situations, and on the basis of the social
interests of groups and social relations between groups in complex social
structures.
It is the task of this second part to spell out some of these social
dimensions of ideologies, and to show why social actors and groups develop
and use ideologies in the first place. Also, we need to study how ideologies
are socially 'invented' and reproduced in society. One crucial component in
this process of reproduction is discourse, which we therefore need to study
separately in the next part, but which as a form of social interaction is
obviously part of the social component of the theory of ideology.
Many traditional and new issues need to be dealt with in this social
framework. Besides the expression of ideologies in discursive interaction,
we must investigate what kind of groups are or may be involved in the
development of ideologies. Second, group relations, and especially those of
power and dominance, and their role in the development of ideologies, must
be investigated. The relevance of'classes' should be assessed as part of such
a broader analysis of group relations. Third, the institutional and organizational dimension of ideologies and their reproduction, such as the role of
politics, education and the media, should be part of a social analysis. And
finally, at the highest or most abstract level, we should explore the role of
culture in the development and reproduction of ideologies.
Again, each of these topics would require a monograph by itself, and
many of these have already been written. However, my approach is more
136
Society
modest. I shall again presuppose most classical work on the social dimensions of ideologies to be known, and set up this part as an integrated
component of a new multidisciplinary framework, which I hope will be
detailed in later theoretical and empirical studies. Also, as suggested before,
I shall not repeat the classical debates, but only examine whether some of
the issues involved will be relevant for my approach or not. For instance,
whether ideologies are essentially'dominarle ideologies or not, is a topic
that will be touched upon only briefly: I have already indicated that I
advocate a broader concept of ideology.
Organizing the social account of ideology
Ideally, this part of the book might be set up so that we start with the microlevel of ideological interaction and gradually extend our scope to more
embracing social structures and processes. However, since we deal with the
fundamental, discursive and interactional, dimension of ideologicál reproduction separately in the next part, this part will generally operate on more
abstract meso- and macro-levels of social structure and culture. Instead of
beginning with the discursive expression and mundane accomplishment of
ideology, this part therefore offers another aspect of the basis and the context
for the study of such discourse, as was also the case for the previous part.
That is, the study of ideological text and talk will later be framed in a
combined cognitive and social account of a theoretical basis that needs to be
established first. If we later want to find out what 'social members' or 'group
members' do or say in a social context, we first need to examine what
ideological membership, groups, group relations, interests, power or dominance mean.
SuCh a decision is partly arbitrary, and an argument coúld be construed for
a different order of analysis. This way of framing the approach also involves
the debate about the micro—macro link that has raged in modem sociology.
Obviously, I cañ t discuss let alone solve all the problems that have been
brought up in this debate. However, the cognitive and discursive componente offer interfaces that have been lacking in this (missing) link. Indeed, as
has been argued before (Chapter 7), the link between groups and individual
persons as social actors or group members, as well as the link between
socially shared cognitions (including ideologies) and actual social practices
of such actors, also has an important cognitive dimension. It is only in their
minds that social actors are able to combine their own, unique, personal and
contextual constraints on ideological practices, with their socially shared
knowledge and opinions about their group membership, about group relations and about social structure.
The society—cognition—discourse link
There can be little doubt, then, that the missing link (also) needs to be
cognitive. Without their socially shared beliefs social actors cannot possibly
know and interactionally accomplish their group membership, which is a
Ideology and society
137
crucial condition for the existence of groups and organizations in the first
place. Thus, even in this chapter we should never forget that it is not the
group, nor the organization, nor any other abstract societal structure that
directly conditions, influences or constrains ideological practices, but the
ways social members subjectively represent, understand or interpret them.
This not only explains the details of the production of discourse and action,
but at the same time allows for the necessary individual variation, deviance,
opposition, dissidence and change, also, of ideologies and other social
structures.
This does not mean that societal structures, groups, power or economic
conditions only exist in the minds of social actors. It has already been
observed that the 'existence' of such social structures is a human construction, and hence a mental as well as a social and practical accomplishment. In
commonsense as well as in theoretical accounts, social structures and
conditions are also postulated to exist independently of the mind, not so
much epistemologically, but simply analytically and sociologically. They
represent another realm of existence, and another level and scope of
analysis, just as physical, chemical, biochemical, biological, physiological,
neurological or cognitive 'realities' exist both as objects of theoretical
analysis and as part of peoplé s mundane experiences.
Thus,,social structures and processes 'exist' both for all practical purposes
and as objects of sociological analysis. They become relevant for interaction
and discourse, however, through the cognitive interface of social actors.
Thus, racism, racist organizations and racist reporting 'exist' for social
actors just as well as their concrete manifestations in everyday racist
discourse and actions. Recognizing the fundamental role of cognition, and
especially of social cognition in such a multidisciplinary account of ideology, does not mean therefore that we 'reduce' the social to the 'cognitive'.
On the contrary, it is theoretically most fruitful to recognize the 'existence' of both, but then to design a theory that integrates these different
dimensions or levels of social reality. In the same way, then, that I have
included a 'social' component in the mind, I now emphasize the important
cognitive dimensions of society. Ideologies, just like knowledge, public
opinion, languages and values, and other socially shared mental phenomena,
may then be studied in a sociological study, even when such a study will
focus more on their 'expression' in typical social 'objects' such as action,
groups or organizations. After the study of social cognition in the previous
part, we thus encounter a cognitive sociology in the present part. 2 The
sociology of knowledge is merely one of the subdisciplines within such a
framework, of which also the sociology of ideology is an inherent part.
Social functions of ideologies
One of the major tasks of such a sociology of ideology is to explain not only
the structures of ideologies as postulated in the previous chapters, but
138
Society
especially also to account for thefunctions of ideologies in society. Virtually
no short definition of ideology will fail to mention that ideologies typically
serve to legitimate power and inequality. Similarly, ideologies are assumed
to conceal, hide or otherwise obfuscate the truth, reality or indeed the
'objective, material conditions of existence' or the interests of social
formations.
Besides such more negative functions of ideology, we may add that
ideologies positively serve to empower dominated groups, to create solidarity, to organize struggle and to sustain opposition. And both at the
negative and the positive side, ideologies serve to protect interests and
resources, whether these are unjust privileges, or minimal conditions of
existence. More neutrally and more generally, then, ideologies simply serve
groups and their members in the organization and management of their
goals, social practices and their whole daily social life. All these functions
are social, and the concepts involved in their description largely sociological. Indeed, they are essentially conditions for the existence and the
reproduction of groups, or for the collective management of the relationships
between groups, rather than functions that only serve individuals. Besides
the cognitive functions of ideologies discussed in the preceding part, we now
may focus on their equally essential social functions.
Racism as example
In order to be able to focus the discussion of the social dimensions of
ideologies, I shall again use racism as the concrete example of a set of
ideologies that have a prominent role in the reproduction of ethnic or
inequality in 'Western' societies. Racism' here will be understood in a
broad, political sense, and involves group prejudices and discrimination
against ethnic or 'racial' minority groups, anti-semitism, ethnocentrism,
xenophobia, and so on. By contrast with many earlier studies of this topic,
racism will not be equated with a racist ideology, but also comprises the
discriminatory practices being enacted on the basis of racist ideologies, as
well as the social structures or institutions involved in the reproduction of
racism, such as political palies, education and the media. In other words,
racism is a complex system of domination, which needs to be analysed at
various levels and domains of society, including those of cognition, discourse, group relations, organizations and culture. 3
Against this background, my examples will focus on the social manifestations and the reproduction of ideologies: Which groups are involved, what
are their relationships, and how for instance are racist or ethnocentric
ideologies 'invented' and spread in white European(ized) societies? What is
the special role of the elites, and of the ideological institutions such as
politics, the media and education? That is, I shall analyse racism to see
ideology 'at work', and especially its conditions and consequences in the
organization of society, and the (dominance) relations between groups,
Ideology and society
139
which will allow us to better understand the societal basis and functions of
ideologies. The next part of this book will then focus on the microsocial
level of the discourses that concretely play a role in the social reproduction
of such ideologies.
15
Group s
Who 'has' an ideology?
After such fundamental questions as what ideologies actually are, and what
they look like, as discussed in the previous part, perhaps the most crucial
question is: who in fact has such ideologies? I have provisionally assumed
that ideologies are essentially social, and shared by groups. ' However, we
have also seen that such an assumption needs qualification: the passengers
on a flight, or the pedestrians waiting for a red light, are not likely to share
one ideology. Indeed, such more or less arbitrarily composed collectivities
might not be called 'groups' in the first place. So, we need to define the
notion of group, and determine which groups typically develop and share an
ideology.
Traditionally, especially in the Marxist tradition, ideologies are of course
associated with the notion of later described in more abstract terms,
such as 'social formations'. 2 More specifically, ideologies were attributed to
the ruling class, which disseminated them to conceal or to legitimate its
power, inequality or the status quo. Similarly, the Gramscian notion of
hegemony implies ideological domination and consent, but also especially in
tercos of a ruling class or power elite, on the one hand, and a large dominated
group of a 'mass public' or simply the citizens whose ideologies are
persuasively inculcated by these elites, on the other hand.
At a later stage, however, with the increasing attention being paid to other
forms of domination, for instance those of gender and 'racé , also other
social groups or formations were attributed ideologies, such as men (or male
chauvinists) versus feminists, or white people (or racists) versus anti-racists.
The same is true for the increasing focus on questions of safety, security,
peace, the environment or various (e.g. sexual) lifestyles, in which also
different groups, collectivities or social movements of some kind are
associated with different positions and ideologies. Peace movements and
ecological movements are just two prominent examples of such 'new
ideological groups, in which the basic principies no longer are of a socioeconomic kind.
In sum, each social group or formation that exercises a form of power or
domination over other groups could be associated with an ideology that
would specifically function as a means to legitimate or conceal such power.
Earlier it was emphasized that also those groups who resist such domination
should have an ideology in order to organize their social practices.
Groups
141
Several of the issues introduced here, such as power, dominance or
hegemony, need to be dealt with later. Here mit need to examine first what
collectivities of social actors may share an ideology, and why.
Groups
Just like various forms of socio-cultural knowledge, and just like natural
languages, ideologies are shared. There are no 'private' ideologies, only
private opinions. Ideologies are acquired, confirmed and changed by social
actors as members of groups, and as a function of the goals and interests of
such groups.
The basic question, then, is what counts as a 'group' in the first place.
Why are the passengers on a specific plane generally not considered a social
group? One reason may be that their membership of the ad hoc collectivity
is simply too ephemeral, and besides travelling safely to the same destination, passengers will not have joint goals and interests. Indeed, they do not
travel as a group, but as individuals who happen to be on the same flight.
Hence, one criterion for groupness may be that collectivities of people must
have some continuity beyond one event.
Of course, the situation is different when some people decide to fly
together, that is, engage in collective action, or when many airline passengers (and not just those on this flight) organize as consumers, that is, as
a group with shared goals and interests, such as safety and service. Similarly,
when their plane is hijacked, the passengers who before merely travelled as
a collection of individuals, may of course become more of a group because
of a shared predicament. Such a shared problem, or a commonfate, in which
people may become mutually dependent, and may want to act collectively to
overcome their plight, may be other criteria for the formation of a group.
More generally, various kind of social conflicts between collectivities of
people typically create groups.
Shared social representations
Note, though, that besides 'objective' social, political or economic problems
shared with others, also cognitive or affective criteria must be involved —
members of a group must know (or believe) about other members, about a
shared problem or conflict, or about possible collective actions. Moreover,
they may share opinions about their common experiences, conflicts or
actions Finally, they have affective feelings of belonging to the group or
about their experiences or activities as group members? In other words, sets
of people constitute groups if and only if, as a collectivity, they share social
representations. 4 For the individual group members this means that part
of their personal identity (self) is now associated with a social identity,
namely, the self-representation of being a member of a social group (see also
Chapter 12).
142
Society
Since social representations take time to develop, and presuppose a
common history of experiences, interaction and discourse, ad hoc collectivities of people do not have such social representations, and hence do not
forro a group according to this definition.
Finally, we might further require that the individual and collective actions
of the group members be monitored by these social representations. That is,
not only should the collectivity, as a set of people, not be ad hoc, but also
the decisions, goals and actions of the members of a collectivity should not
happen to coincide or be similar, as was the case for the individual
passengers travelling aboard the same flight to the same destination. Thus,
group members act as group members when these actions are (also, though
not exclusively) based on shared knowledge, attitudes, ideologies, norms or
values (see Chapter 3).
Thus, slightly less ephemeral than the group of passengers on an airliner,
we may take the example of a demonstration. Here, membership is not
arbitrary, but members share opinions, and at least one goal. They do
something together, namely, protest against some social situation, action or
policy they disapprove of, and they know it (also about each other).
However, although such a protest and the opinions that give rise to it may
well be ideological, a protest demonstration — as such — need not be an
ideological group either. The goals and opinions shared by the demonstrators
as well as the collective action, after all, may be strictly contextual, and not
go beyond the occasion at hand.
On the other hand, some demonstrations may be based on shared group
attitudes and ideologies, for example a demonstration of environmentalists
against dumping nuclear waste, or an anti-racist demonstration against a
racist panty. In that case, the attitudes and ideology are shared by a broader
group than just the participants in the demonstration. Members of the
demonstration in that case are a subgroup of a larger group, such as a social
movement, and the protest one specific manifestation of ideologically based
attitudes.
From this theoretical analysis, as well as from these examples, we may
conclude that ideologies and groupness mutually seem to define each other.
On1y groups may develop ideologies, and the definition of groups in turn
presupposes not only shared social conditions, experiences or actions, but
especially also shared social representations, including ideologies.
This circularity of the definitions of ideology and groups is both apparent
and theoretically welcome in a theory of ideology. First, although all
ideologies are group-based, not all groups need to develop an underlying
ideology. Shared knowledge and some shared group opinions may be
enough for many forros of collective actions and goals, as would be the case
for our example of a group of people regularly taking their vacations
together. On the other hand, many groups (or maybe social groups stricto
sensu) can only reproduce themselves, and continue to exist, if they — or
their members — satisfy a number of social criteria, including access to
specific resources, as we shall see in more detall below. Some of these
Groups
143
resources are not just material, but also symbolic, (knowledge, information,
education, status, etc.), as is the case for politicians, professors and joumalists, among many others. Since such symbolic resources are defined in tercos
of socially shared representations that actually define their social value, we
again are back at the socio-cognitive level in order to define groups.
Moreover, many groups are primarily defined in terms of these social
representations (e.g. opinions, ideologies) themselves, as is the case for
Christians, socialists, feminists, anti-racists or peace activists and many
other social movements.
And finally, even for groups that seem to be constituted also or primarily
in terms of material resources (like the poor and the rich, the homeless and
the unemployed), we have seen that socio-economic conditions are relevant
for the group only if their experience is shared, and hence framed in tercos of
shared knowledge or beliefs, that is, if group members actually feel and
represent themselves as members of such a group; or conversely, if they are
represented by members of other groups as such, and are treated accordingly. And finally, for most groups continuity and reproduction presupposes
either individual acts of social actors as group members, or collective action,
which in both cases presupposes shared social representations of the
members.
Note that this does not mean that being poor or homeless is 'all in the
mind' , and that socio-economic conditions are thus reduced to their mental
representations. Of course not. But for someone who is poor or homeless to
feel and represent him/herself as a member of a group, and not just out-ofluck as an individual poor or homeless person, such economic conditions
still need to be interpreted and especially also represented as being shared by
others.
Nor does this argument imply that groups are only constituted by social
representations. They are also characterized of course by their (lack of)
access to material or symbolic resources,.by collective action, discourse and
other social practices. However, whatever the 'objectivé socio-economic
base of a collectivity of people, they can only constitute a group if they share
social representations that give collective meaning to these social circumstances. It is also in this sense that groups are not merely a societal construct,
but also constituted mentally through shared representation. Groups are also
constituted by their members, as well as members of other groups, by
feelings of belongingness, shared memories of collective experiences, and
more generally by social representations, or precisely by the fact that others
do not share these representations or challenge them. And if groups should
be defined by the social practices of their members, the same necessary
precondition obtains, as we have seen — social actors can only act as group
members if they develop and share such social representations in the first
place.
If groups are constituted by the shared social representations of their
members, but not all groups have ideologies, we must later formulate farther
144
Society
conditions on'which groups under what conditions actually develop ideologies. Thus, whereas a group of vacationers, as such, may not (need to) do
so, battered women, managers, or peace activists are more likely to develop
some forro of ideology. These conditions might be socio-cognitive, for
example when the specific social representations of a group need further
organization and foundation, or when group members need to co-ordinate
their actions or engage in collective action. And they may be socio-cultural
or political, namely, for example for effective group reproduction, organization, sanctions of norms, domination of other groups, conflict resolution, and
more generally the effective realization of its goals. Below I shall return to
these further conditions on the development of ideologies.
Social categories versus social groups
The criteria of group continuity and social identification typically apply to
social categories of people, defined in terms of more or less permanent
properties, such as age, gender, ethnicity, origin, class, language,
religion, sexual orientation or profession. Hence, women and men, white and
black, young and old, and poor and rich may well develop ideologies that are
related to the position and the interests of the members of this category in
society.
However, general social categories are again too broad to form groups as
defined aboye. After all, it is not likely that all women, or all rich people,
share the same overall ideology, even when they share similar social
experiences or act similarly in certain social situations. Taking the example
of class struggle, feminism, or the civil rights movement, we see that these
apply to groups of people who belong to a social category, but also share
specific goals, norms, values, and in general some forro of awareness about
these. And this awareness or group feeling was defined as social identity and
hence as a form of shared social (selprepresentation. Social movements may
defend the interests of all workers, women or blacks, but as groups they have
more specific goals and interests that need not be shared by all members of
their respective social categories. 5
Equally general are those collectivities of people that are precisely defined
by their ideologies, such as liberals and conservatives. Again, the question
may be raised whether these are 'groups' in a more narrow sense — do all
conservative people in the world form a group? Should their ideological
stance be taken as a more or less permanent property, as is the case for
gender, age or ethnicity? It may be assumed that members of such 'groups'
identify more or less strongly with them, precisely for ideological reasons. If
shared social identity is a sufficient criterion for the definition of groups,
then this collectivity of people would constitute a group. They may even
have some overall goals. On the other hand, unlike demonstrators or
members of social movements, members of such ideological categories do
not, as such, participate in joint activities, but at most in similar activities,
Groups
145
as voting and engaging in liberal or conservative actions and discourse.
is why the analysis in Chapter 28 of a concrete example suggests that
servativism' be considered as a 'meta-ideology' that controls dimens of the ideologies (e.g. those of neo-liberalism, sexism or racism),
r than as a proper group ideology.
iother general type of group is based on profession. Doctors, nurses,
^ ssors, journalists or carpenters thus may form a professional group,
h obviously has similar activities, goals and interests, and with which
i or most members may identify. Such groups have professional values
lorms, and opinions and attitudes about professional practices, as well
oup-specific expert knowledge. Though possibly universal (specialized
^ ssions exist in virtually all societies and cultures), this type of group
is like a plausible candidate for the development of group ideologies,
. Y cially given the relevance of conflicting interests between differein
professions. But again, members of the same profession across the globe
only rarely engage in joint activities, although some do, for instance at
international conferences.
Between these very general (if not universal) categories of social actors on
the one hand, and the fleeting group membership of a demonstration or a
team, we have the groups of people that constitute organizations and
institutions, such as political parties, parliaments, universities, unions and
corporate businesses. Again, identification with such organizations defined
as 'groups' is plausible, and there are shared (and even joint) activities, goals
and values, as well as similar interests. Note, though, that there is a problem
here: As individual institutions or organizations they may no more have their
own specific ideologies than their members. We do not speak of 'the'
ideology of a specific union, but rather about a union ideology in general.
Similarly, in business corporations we may find more general corporate
ideologies (or variations of them), and not so much the ideology of one
specific business corporation. If such corporations are large, as is the case
for multinationals like IBM, however, a common 'culture' may develop, and
such a culture of shared norms, values and goals might in a way be identified
as the corporate 'ideology' . 6
Another criterion, maybe a decisive one, for the definition of the social
group basis of ideologies, is social conflict, struggle or any other kind of
interest-based opposition between groups, wheiher over material or over
symbolic resources. This is traditionally the case for classes and class
struggle, and in Marxism, obviously, ideologies were primarily related to
groups such as workers and 'capitalists'. The same is true for feminists
versus male chauvinista, or anti-racists against racists, and so on. In such
cases the dominant groups will tend to develop an ideology that serves the
reproduction of its domination, and the dominated groups may develop an
ideology as a basis for its attitudes, opinions, practices and discourses of
resistance or opposition. Membership, activities, goals, social position,
values and group resources are al rather easily identifiable here, and given
146
Society
these as basic categories for the definition of ideological schemata, these
could be the prototypical ideological groups. Other groups (say, a category
like 'women', or a profession like 'doctors', or a party like the ChristianDemocrats) would generally be defined by only one or a few of diese
categories . If we define ideologies in tercos of their social functions (see below), then
sharing beliefs, the co-ordination of social action and interaction, providing
identification, common goals, organization and in general defending group
interests, are major conditions for the constitution of ideological groups.
Collectivities of people as defined by one or more properties (such as age,
profession, goals, income level, political orientation, etc.) thus will tend to
be more like ideological groups if these ideological functions apply to them.
We need a detailed sociological theory of social groups in order to be able to
make such criteria explicit.
One such criterion may also be the degree of institutionalization. This first
of all excludes all ephemeral groups, such as plane passengers and participants in a demonstration. It also eliminates general social categories, such as
socio-biological ones like men and women, blacks and whites, old and
young, or socio-economic ones like rich and poor, or the unemployed. As
suggested, these general categories may well be the broader collectivities
from which more specific ideological groups are recruited, however, as is the
case for feminists as members of the category of women. Many ideological
groups, such as feminists, socialista, environmentalists, anti-abortionists and
so on, are not merely defined by shared identities, goals, positions or
resources, but also by the fact that they tend to organize in institutions, such
as parties, non-governmental organizations, Churches, sects, and so on. They
often have explicit, self-styled or elected leaders or officers, headquarters,
membership fees, publications, meetings and so on. Such institutionalization
may play a prominent role in recruiting new members, setting goals,
formulating norms and principles (and indeed ideologies), securing resources, and especially the co-ordination and effective execution of actions that
realize the goal of the organized group.
We may conclude this discussion by assuming that there cannot be a clear
and explicit boundary between social groups in the more specific sense and
any other collective of people defined by one or more shared characteristics.
Generally, however, I shall assume that a social group must be more or less
permanent, more or less organized or institutionalized, and reproduced by
recruiting members on the basis of identification on a specific, more or less
permanent set of properties (like gender or income), shared activities and/or
goals, norms and values, resources, and a specific position (often of
competition or conflict) in relation to other social groups. Groups that satisfy
most of diese conditions will then be assumed to be most likely to develop
shared ideologies that will serve as the basis for organization of the actions
and cognitions of their members in such a way that the aims of the group are
optimally realized.
•1
Groups
147
Groups versus members
There is one thorny theoretical problem we need to address here, and that is
the specific, emergent nature of a group as being distinct from the set
constituted by its members. Throughout this study many observations are
made about the ideologies and other mental or social properties shared by a
group. We have assumed, for instante, that journalists as a group develop a
professional ideology and that other collectivities of people may do likewise
in specific social circumstances.
However, the problem is that we would like this also to be true when one
or a few individual journalists do not share such an ideology. Indeed, many
groups may have ideological 'deviants' or 'dissidents'. If this is the case, the
notion of group may at least sornetirnes be distinct from the set of its
individual members. Maybe 'groupness' only requires that most or many -of
the members share some property. However, such fuzzy criteria also make
groups 'rather like fuzzy sets rather than strict sets of members. Indeed, as
with sets, groups may theoretically exist if they do not (as yet, or any longer)
have any members at all!
Apart from the set theoretical and quantitative dimensions of groups, we
may also ask whether groups have emergent properties that (sets of)
members do not necessarily have. Indeed, are there mental representations
(like knowledge and ideology), collective actions or group relations that do
apply to the group, but not necessarily to (all of) its members? Obviously,
this is the case. As we shall see in the next chapter, social group relations
such as power and dominance are defined for the group as a whole, and do
not necessarily apply to all members. Indeed, despite male dominance in
society, not all men are dominant with respect to all women they interact
with, flor do most men act dominantly against women all the time. Groups
may similarly have a collective past, history and experiences that not all
members personally have, as is typically the case for the Holocaust for the
Jews. From this example it is but one step to collective memories and hence
to shared social representations: Jews as a group have social representations
about the Holocaust and anti-semitism, although there may be individual
Jews who doñ t. These few examples suggest that indeed groups may have
attributes that are not necessarily those of (all) their members.
It is likely that the same is true for ideologies. That is, because of their
history, collective experiences, social position, and social relations with
other groups, groups may develop and reproduce a specific ideology. Like
'having a language', thus, 'sharing an ideology' is such a property that
should be defined at the societal level, that is, for the group as a whole. In
the same way as a social group is an abstraction, or an ideal type, also
ideologies may thus be seen as an abstract property, much like languages
such as English or Chinese are abstract systems, at least at one level of
analysis. Such a system is not simply the same as the actual language use of
all speakers of English or Chinese. Indeed, there are languages that, as
linguistic systems, have survived their users. Similarly, socialism as an
148
Society
ideology will still be an ideology even when the last socialist has switched
off the light.
The macro—micro problem
These observations, however, require a further analysis of the relations
between social abstractions, systems, collective properties, and groups on
the one hand, and group members as actual people — as well as their minds
and actions — on the other hand. This is an example of the well-known
macro—micro problem in the social sciences. 8 In the same way as the system
of the Chinese language must at least be partially known in order to be
'used' by concrete speakers, we may assume that a similar condition holds
for the role of ideology in the monitoring of social practices in general, and
of discourse in particular. That is, if ideologies are defined only for groups,
if groupness presupposes shared social representations (or a social identity),
if social representations are mental, and if groups as such doñ t have minds,
then we must assume that groups can only 'have' an ideology if at least a
qualified number of their members share at least part of such an ideology.
Now, what exactly does that mean?
One trivial answer already formulated aboye is simply quantitative. That
is, a group an ideology if most of its members share most of the
propositions that define such an ideology, where the fuzzy quantifier 'most'
may be assigned any value between, say, seventy-five per cent and 100 per
cent.
Somewhat less trivial would be to replace the quantifier for the number of
propositions by the set of'essential' or 'core' ideological propositions,
namely, those that are the specific, defining or prototypical fundamental
beliefs of a group. For instance, people would not qualify as neo-liberals if
they did not share the core ideological propositions based on the freedom of
the market. This is relatively straightforward of course for groups defined
primarily by their ideologies.
But what about journalists? Does a journalist who does not believe in the
core propositions based on the value of the freedom of the press, exclude
hico- or herself from the journalistic ideology, and hence from group
identification? Would such a journalist not be like a prototypical journalist,
and hence be defined (or define himself or herself) as a relative outsider, and
actually be marginalized, as we may indeed observe in journalistic (or other
professional) practice?
Are the well-known forces of conformity, including socialization, schooling, the media, sanctions, marginalization and other social practices to
enforce ideological alignment of members, a social manifestation of the
necessity to defend at least the adherence to a core of ideological propositions by all members? Such seems indeed to be the case. Again, the
comparison with language (grammar) may be instructive: use is not merely
regulated by mutual intelligibility but also by socially enforced standards of
minimal correctness for many social situations, such as schooling and
Groups
149
or keeping a job. Personal variation is possible, but some normative
animar needs to be respected in specific social situations.
least a minimal ideological core should be respected 'in' the group,
e still need to qualify by how many or which members this needs to
case. Again, we may use a qualitative criterion, namely, the 'core
;rs', such as leaders, the elites, all people with responsibilities, and in
1, thus, the 'ideologues' of any group. This is socially necessary for
>up, in the first place, in order to ideologically reproduce itself. At
eme members need to teach the ideology to newcomers or new
ions. Second, at least some members need to monitor social practices
rice the applications of the ideology by current members. Third, at
)me members need to be able to reformulate and adapt the group
,y to new social developments, circumstances or changed relations to
roups. Fourth, at least some members need to be able to formulate
tributé (fragments of the) ideology throughout the group. These and
deological core activities need to be adequately carried out for any
to reproduce its ideology and the social practices and social position
basen on it. In other words, we may conclude again, and again rather
vaguely, that ideological reproduction presupposes at least a core of elites or
ideologues to perform these functions.
Of course, for different groups or institutions such ideological activities
may vary considerably: the Catholic Church does this in a different way
from the feminist movement or the peace movement. Also, the conditions on
either the number of ideological members or the number of ideological
propositions to be shared by them, may be very different for different
groups. Traditionally, in the Catholic Church one might be excommunicated
for adhering to a specific heresy. Something similar might happen for a
strictly ideological political party or specific social movements. In some
(extreme) cases all members need to ascribe to all ideological propositions,
whereas in others only to a core of basic ideological principies, or again,
only a core group of people need to kiiow all or most or only the core
principies. But if only a small group knows and shares only a fragment of
the (original) ideology of a group, so that full ideological reproduction
becomes impossible among newcomers, we may expect ideological decline,
or ideological change, or indeed the resolution of a group. Since, however,
ideologies may often be written down in explicit textbooks, Bibles, catechisms, histories of movements, party programmes, corporate 'rnission
statements', organizational statutes, and similar ideological writings by
ideologues, there is always a possibility that at least some group members
are able to keep the ideological Eire burning for a long time.
What is sharing?
Finally, there is another aspect that needs to be examined when we study the
relations between the ideological group and its members, namely, the precise
social and cognitive status of sharing. Aboye we have seen what kind of
150
Society
social dimensions may be involved, namely, how many, and what kind of
members need to share how much of an ideology. Now, the question is what
this actual 'sharing' means — identical'copies' of propositions in the minds
of the relevant members, as when computers run copies of the same
program, even when different personal uses of the program are made?
Again, the comparison with grammars may be instructive. In order to use the
language more or less grammatically, we may assume that language users
need to acquire more or less the same rules of a grammar. Of course there
may be personal variation, for example due to schooling and other forms of
learning, in the amount of mies acquired or how the rules are applied. That
is, despite such variation, most speakers of the language must have more or
less similar copies of the core rules of the same grammar.
The same, we assume, should be true for the basic social representations
of a group, that is, its ideology. On many occasions, such ideological
principies may actually be formulated, for example in contexts of admission,
inclusion, socialization, initiation, teaching, jurisprudence, sanctioning, marginalization and exclusion. Of course, such formulation takes places in
variable discourses, and not directly in tercos of (abstract) ideological
propositions, so that acquisition, even in ideal cases, is often less strict than
acquiring the rules of grammar. However, as with other social principies,
such as norms and values, there are many social practices and discourses
expressing or enacting the underlying ideological principies, so that by
continuous repetition and experience, some fairly similar ideology fragments
will be acquired. Again, this will be highly variable for different ideological
groups. Also, it need not be emphasized that the knowledge that group
members have of such ideological propositions need not always be explicit
or even conscious (as is also the case for the rules of grammar), if only they
are able to apply them more or less adequately.
In sum, for the moment, we have no theoretical alternative but to assume
that a group an ideology if at least some (or most, depending on the
group) members share at least some core (or most) ideological propositions.
Sharing, in this case, means that these members have fairly similar propositions stored in their social memory. In other words, if an ideology is taken as
an abstract system of the group as a whole, it is concretely (mentally)
'distributed' over its members. That such members will make (sometimes
vastly) different uses of this ideological system in different social contexts,
is obvious, and defines the large variation in ideological discourses and other
social practices. We shall come back to such personal and contextual
variation later.
Multiple identities and conflicting ideologies
As has been suggested several times aboye, individual social actors may be
members of various social groups, each of which may have its own
ideology. This is one of the fundamental reasons why the expression of
ideologies by such actors in specific situations may appear incoherent or
1
Groups
151
oven inconsistent. The point here is that since different groups may have
different interests (membership devices, activities, goals, norms or resources), also their ideologies, which cognitively represent such basic interests,
may be in conflict in the decision 'how to speak or act' in a specific
situation. Depending on the situation, one identity and hence one ideology
may be more relevant or more important, so that strategic choices can be
made in the management of conflicting beliefs and interests. We already
encountered the prototype of a middle-class female black journalist, who
will probably let her journalistic ideologies and practices prevail over those
suggested by the other group ideologies, at least if she wants to keep her
lob.
Only sometimes can such group ideologies and practices be combined.
Middle-class ideology will often combine well with the middle-class bias of
the media, most sources, most news actors and most of the public. No
conflict of interest will be very likely here. As a woman, the black journalist
may be partly accepted (or even forced) to 'behave' as a joumalist, although
she may get story assignments with a womeñ s perspective (but seldom a
radical feminist one). This will be even less the case for her as a black
person, although in times of ethnic or racial conflict and crises, 'ethnic'
stories may be assigned to her. But in general the social rule is: believe and
act like most of us in 'our' group. Transgressions of the rule, and outright
deviance and dissidence will be sanctioned by marginalization, exclusion or
elimination, whether physical, economic, social or cultural.'
Group categories and membership
Ideologies have been assumed to be organized by a group schema consisting
of a number of fundamental categories that codify the ways people define
themselves and others as group members. These categories have mental
aspects, but also social ones. Thus, whereas membership may be construed
as the mental representation of the relations individuals have with social
groups or categories, it also needs to be accounted for in more sociological
terms. Thus, it may not be sufflcient that group members consider themselves to be members of a group. It is also important how others perceive
them as such. Indeed, individuals may 'naturally' belong to, and be
considered and accepted, as members of groups or categories, as is the case
for women, children, or whites and blacks, but in many other groups, the
process of admission and recognition follows a more complex social
process. This also affects the role of membership in the reproduction of
ideology.
Besides the 'natural' categories just mentioned (which obviously are
social constructions in their own right), membership should first be examined for those social groups people may be 'born into', and to which they
thus have more or less involuntary access. Class and caste are the most
prominent example of such groups, and at the same time the classical
152
Society
example for the development of ideologies. Although in later life people
may well 'change' class or caste, they are assumed to be class members as
long as they are unable or unwilling to change such membership.
Class membership is as complex as the very notion of class and socially
constructed in terms of much more than just socio-economic parameters,
such as family income, occupation or position. Various types of nonmaterial, symbolic 'capital' may be indicative of oné s class, such as status,
respect, accent and language use, knowledge of the arts, and so on for the
upper and upper middle class, or precisely the relative (real or attributed)
lack of such symbolic resources for the lower classes. Changing class,
especially upwardly, therefore also requires more than just a change of
material resources, as is shown by the negative categorization of nouveaux
riches as not really belonging to the upper classes. Indeed, specific upper
classes (e.g. of nobility) cannot be entered otherwise than by birth. For
others, both material and symbolic capital, such as a good education, a good
school, and acquired 'culture' may be needed to access the higher class if
they did not acquire such resources as lorñ class members. 11
Both for established members as well as for new members, class member:
ship is also associated with ideologies. Indeed, the very reproduction of the
socio-economic interests of the class, including both its material and
symbolic resources, is one of the main functions of class ideologies. That is,
when matched with the structure of ideologies, class provides a 'membership
device' that is essentially resource-based: people define themselves and are
categorized, recognized or admitted by other members primarily in terms of
a specific set of (socio-economic and cultural) resources. For the successful
reproduction of the class, therefore, group members learn either from birth,
or as newcomers, the ideology that aliows the protection of these resourcebased interests. Economically, this may mean opposition against various
forms of wealth and income distribution, high taxes for the rich, and so 9n.
Symbolically, this may mean exclusive or preferential access to special
schools, clubs, professions and forms of culture. Ideologically, such
privileges will tend to be legitimated by claims of natural or social 'rights'
(birth, marriage, heritage) and/or merit (hand work, learning).
For professional groups, membership and access are usually well defined
in terms of legal or traditional membership criteria, such as education,
degrees, diplomas and expertise. Lawyers, doctors and professors can
become such only when being officially evaluated and qualified, whereas for
journalists less strict qualifications may be needed. Membership of such
groups is usually based primarily on type of activity and expertise (advising
clients, healing patients or teaching students). The interests of such groups
are also tied to specific, symbolic resources such as legal, medical or
scientific knowledge and expertise, as well as status and respect accorded the
professions in a particular society.
Given their nature, we may expect that professional groups develop
ideologies especially as a function of the interests tied to their activities and
their special resources. Thus, freedom of the press, independence of the
1
Groups
153
courts, autonomy of the universities, as well as freedom of information and
science, are well-known elements of the basic ideological beliefs that reflect
such interests. Also for these reasons, membership is rather strictly regulated; the number of people having access to these resources should be kept
fairly small, so as to maintain the economic value of professional services
and expertise. Most professionals oppose lowering standards' or 'mass
universities' and insist on a self-regulated threshold of access, such as
special exams, in-house training or specialization. Successful reproduction
of the group through the protection of these special interests thus also needs
to be articulated in various professional ideologies, for instance about the
importance, relevance or the functions of these professions (serving justice,
health care, knowledge and education, or informing the public). Similarly, in
order to protect such interests, the activities of the professional may also
internally be judged on the basis of ideological values (justice, truth,
reliability, fairness, etc.)."
Groups and their membership may be constituted also on the basis of their
social goals, usually in relation to their norms and values, as is typically the
case for various social movements. Feminists thus form a group typically on
the basis of their goal to end male domination and gender inequality. The
essential values involved in their activities for the realization of that goal are,
for example equality, independence and autonomy. The same is true for
socialists, environmentalists, human rights activists and similar action or
advocacy groups, on the one hand, and for nationalists, racists, and antiabortionists, on the other hand. Membership criteria in Chis case will
therefore be personal choice, ideological alignment and recognized activities
that contribute to the realization of the common goal. As is the case for other
groups, ideologies of these goal-defined groups reflect their main interests,
such as gender equality or ethnic autonomy. At a higher level of abstraction,
the same membership criteria and ideological development apply to ideological groups, such as conservatives and liberals. In Chis case, the main
13
membership criterion is precisely the ideology itself.
Although most groups and their identity are defined in terms of their
relations to other groups (outgroups), some groups are specifically defined in
terms of the social position of their members within the group. This is the
case for leaders, managers, chiefs and in general the elites, in relation to
subordinates, underlings, ordinary people, the masses, citizens, the 'people'
and so on. That is, apart from hierarchical position, their main resource is
power. Membership conditions and criteria may in this case be appointment,
election and self-selection. Ideologies of such groups should be primarily
articulated in the interest of the reproduction and the legitimation of their
crucial resource, namelr, power, as is typically the case for leading
politicians and managers!
Although not complete, this categorization of various groups and their
membership criteria shows that there is a close interplay between ideological
categories and the essential dimensions of social access, membership,
154
Society
activities, goals and resources of groups. Ideological structures have precisely been postulated as the cognitive reconstruction of the main social
conditions for the existence and reproduction of various social groups. In
other words, the essential conditions of existence, organization, reproduction
and the social practices of groups and their members have both social
dimensions and mental ones. Here, ideologies and groups mutually constitute each other. No group can socially exist and act without a group
identity and shared ideological behefs of its members. Conversely, no group
ideology will develop unless collectivities of people start to act, co-ordinate
and organize as a group. Indeed, a large part of the social practices of many
groups, and especially of teaching, communication and discourse, is precisely geared towards the development of a common ideology.
In sum, in rather general terms, social groups and their members may be
distinguished by
• who they are, as defined by more or less permanent characteristics, such
as gender, ethnicity, caste, class, age, religion, language or
origin;
• what they do, as is the case for professionals;
• what they want, as is typically the case for advocacy groups;
• what they believe, as is true for advocacy groups, as well as for religious
and ideological groups such as conservatives and progressives;
• where they stand, as for all groups defined in tercos of social position and
their relations to other groups;
• what they have or don 't have, as for all groups whose identity is based
primarily on the special access or lack of access to social (material or
symbolic) resources, as is the case for the rich and the poor, the
employed and the unemployed, the homeless and the home-owners
the famous and the infamous, the educated and the non-educated, the
intellectuals and the non-intellectuals, and so on.
The categories that define this typology of groups are intentionally the
same as those that form ideological schemata (see Chapter 5). That is, our
approach precisely emphasizes the mutual constitution of the social and the
cognitive dimensions of groups. Most social criteria discussed aboye for the
constitution of social groups can thus be articulated in terms of categories
that also organize the social cognitions shared by group members.
Ideologies without groups?
The assumption of the mutual constitution of groups and their ideologies
raises an important final question: are groups necessary as the 'social basis'
of ideologies, or would it be more appropriate, at least in some cases, to
allow ideologies to 'exist' more independently? There can be litde doubt that
there are collectivities of social actors that can be defined in terms of nonideological social conditions, as is the case for socio-economically defined
groups (classes) or professions. That such groups also need to share social
155
Groups
beliefs and ideologies in order to co-ordinate the actions of their members
and to reproduce thernselves, has been shown aboye.
But what about groups of which membership is much more fuzzy and
primarily ideological, such as the feminist movement or the peace movement?
Can we simply say that afi 'members' of the feminist movements are feminists,
and that those who are not members are not feminists? What about women
who share some tenets of feminism, but not others, or those wornen who do
share many ferninist propositions, but do not consider themselves feminists? Is
the feminist movement a well-defined group in the first place? Or should it be
defined in tercos of a fuzzy set, in which some wornen may be 'more or less'
members, depending on the amount of their ferninist beliefs, or their degree of
identification? Or should we use some version of prototype theory and
distinguish between more or less prototypical feminists?
In other words, especially for social movements that have a more
individualistic orientation, it might not be an already existing group which
has' an ideology, or an ideology that requires a group, but individual social
members who adopt, to a greater or les ser degree, ideas of an ideology. Such
an ideology would then rather have the status of any other system of ideas,
such as a philosophy or a theory, and could have been developed by one or
more individuals, whether or not it is being shared or adopted by many
people or a collectivity of social actors.
Such a more individualist approach to certain ideologies would at least
avoid the theoretical problems associated with the definition of ideological
groups as indicated aboye. It would account, by definition, for the vast
individual differences in the adherence to certain ideological propositions,
and for a more dynamic process of ideological change and renewal. It would
emphasize peoplé s individual decisions in adopting ideology fragments,
and explain the personal variation in the enactment of ideologies in everyday
social practices. We would not peed to worry whether a conservative
ideology, for instance, is shared by a group, but may simply say that given
such an ideology as a socio-historical phenomenon individual social members may espouse one or more of its tenets, but may reject others.
In this framework, people do not become all-or-none members of, for
example, the 'club of conservatives', but simply use (fragments of) a set of
beliefs as a resource in the organization of their own knowledge and
opinions and the social practices based on these. Ideologies of this kind
would be like 'personal organizers' rather than social (group) organizers.
Indeed, this approach would also account for the seemingly curious situation, signalled aboye, that we may have ideologies that have no 'members',
or rather adherents, at all, as we also have abolished religions or theories.
Ontologically, ideologies like this would only exist as a form of (possibly
specialized) historical knowledge, or expressed historical documents or
treatises, but no one would still lelieve in' them.
These argumenta for a more individualistic approach to (at least some
types of) ideology are quite persuasive. They again suggest that an exclusively sociological definition of ideologies is incomplete. The processes as
'S
156
Society
described are fully accounted for in the cognitive theories presented in Part
I. Indeed, personal beliefs, experiences and practices have been shown to be
associated with specific or general mental models, as well as by other
representations in episodic memory. Individual social actors may thus
flexibly adopt and personally integrate whatever 'ideas' there are available
in the public domain They may for the same reason more or less identify
with one or more ideologies or social movements, or may recombine
elements from several ideologies. Women may experience and thus interpret
male practices as women, and bring to bear shared social representations of
women as a category, but not necessarily identify with feminism as a social
movement or interpret their experiences in tercos of a feminist ideology. The
same is probably true for many religions, political convictions, lifestyle
conceptions, and on any other system of attitudes (e.g. about abortion,
nuclear energy, or the environment).
Paying due allowance to this individual dimension of ideology, or rather
of the 'uses' of ideology, however, does not mean that the social, collective
dimension can be simply dispensed with, for the many reasons given
throughout this book. Thus, the feminist and the peace movement do not
merely consist of sets of like-minded individuals. First of all, at the social
side, there is social interaction between such individuals, and some of there
interactions are engaged in as a consequence of, or precisely as a condition
of, sharing specific beliefs, that is, by social actors as lelievers'. Second,
social movements are also defined in tercos of collective actions, such as
demonstrations or strikes. Third, they have many forros of organization and
institutionalization; they have leaders, programrnes, socio-economic resources, and so on. That is, they may have all the characteristics that define a
group.
What a theory of ideology needs to explain, then, is precisely the
dynamics that relate social members to ideologies and to the collectivities
that are constituted by shared experiences, beliefs and ideologies. We need
to know how individual membership, identification, allegiance, solidarity
and active participation are being defined in relation to such a collectivity
and its organization. It should be examined how groups may grow and decay
as a result of the actions and participation of individuals. That ideological
'groups' may be quite loose or fuzzy in their definition, and their membership or adherence flexibly defined in tercos of the interplay between personal
beliefs and socially shared beliefs, does not mean that we can dispense with
the social dimension of ideologies in terms of groups or similar collectivities. It is this interface between the individual and the group that is one of
the theoretical problems that need to be examined in a theory of ideology.
Racist groups
This dilemma about ideologies and their relationships to groups and individuals is particular clear in the study of racism. Indeed, which 'group' has a
1
Groups
157
racist ideology? It would be easiest call this group the 'racists'. However,
the delimitation of that group would require the definition that racists are all
people who share a racist ideology, but, if not circular, this would be rather
trivial. Moreover, as we have seen aboye, many people may share some
social opinions (prejudices) based on racist ideologies, but need not share a
fully fledged racist ideology.
One could also define 'racists' in terms of their organizations, for example
by identifying them as members of racist organizations, but the circularity or
triviality would in that case also hold for the definition of such organizations. Moreover, it is obvious that there are more 'racist people' than people
who are members of racist parties or organizations.
Another possibility would be to identify the group with all white (European) people. But that is obviously inadequate if we assume that racism is
not an inherent or essential property of white Europeans. Indeed, some
whites share an explicitly anti-racist ideology, whereas some members of
minority groups may support racist. ideologies.
Moreover, we have already seen that it is pointless to distinguish sharply
between those (whites) who are racists and those who are not. Rather, we
must assume that elements of racist ideologies and attitudes are distributed
unevenly over the white group: sorne people will only share sorne or
moderate racist beliefs, whereas others have many and blatant racist beliefs.
Indeed, the same is true for anti-racist ideologies.
Instead of distinguishing between racists, non-racists or anti-racists,
therefore, it is much more adequate to speak of racistpractices. Practices
may then be called 'racist' if they contribute, more or less directly, to ethnic
or racial inequality. Racist practices (and not just any unethical or unacceptable activity) are such also because of underlying opinions, attitudes and
ideologies, for instance those that imply any forro of non-egalitarian
relationships between dominant and dominated ethnic groups.
The example of racism shows that the association of ideologies with
social groups is not a straightforward matter. We might say that managers,
joumalists or other more or less easily definable groups share a professional,
occupational or other ideology. But other categories and groups are much
less well defined, even if they share an ideology, or especially if all that
defines them is their ideology and the social practices derived from them.
The same is true for feminists, environmentalists, and more generally for
ideological groups, such as progressives, liberals or conservatives. They
forro 'groups' of a very different nature than for instance a specific action
group or profession, and are much more distributed over other groups, both
socially, regionally and internationally. Sometimes they are órganized, as in
liberal or conservative political parties, sometimes they forro sects or
churches, and sometimes also an international movement, as is true for antiabortionists or environmentalists.
Most people who have racist beliefs and act and speak accordingly, will
vehemently deny that they are racists. That label is officially sanctioned as a
negative qualification, as is that of being a bigot or intolerant in societies
158
Society
where lolerance', 'equality' and 'democracy' are dominant official (ideological) values.' Hence racism is a typical example of an ideology applying
to a group as it is defined by others.
The provisional conclusion from this brief analysis must be that the notion
of group needs to be taken in a broad sense in order to be able to associate
ideologies with groups. Crucial, as we have seen, are
1 the development and sharing of social representations;
2 identification of members with the group;
3 the defence of specific resources (such as citizenship or equal rights in
all domains);
4 relations to other groups (e.g. resentment against immigrants);
5 specific activities (such as discrimination) and at least a vaguely shared
goal (segregation, immigration restriction, etc.).
Social members who identify with these criteria are, by definition, group
members, but the boundaries of the group are ill-defined. As is true for
cognitive category theories of prototypes, thus, we may have more or less
prototypical 'racists' and 'anti-racists'. Skinheads who beat up Turkish
women only because they are Turkish are more typical for the commonsense
notion of racists than are cabinet ministers advocating immigration restrictions or professors who have less confidence in black female PhD candidates
than in white male ones.
These examples again show that ideologies as well as social groups and
social relations (and their self- and other-perception) are all social constructs, which both have cognitive and social (societal) conditions and
consequences. The distinction between 'cognition' and 'society' in this case
becomes purely analytical and theoretical. In the everyday life of group
members who participate in a group and its ideology, these cognitive and
social conditions and criteria are inextricably interwoven — one may
(socially) 'be' a Christian simply by 'defining' oneself as such.
Very often actions and interactions are required if one is to socially
display or prove oné s membership. However, in that case the sociocognitive definition or construction also applies to these actions themselves.
Doing or saying something also needs to be (mentally) planned or interpreted'as' feminist, and will not 'inherently' be so. As we have seen already
in the chapter on identity (Chapter 12), while acting'as a group member' it
is not merely the action itself that identifies a member, but rather the specific
meaning attached to that action. Hitting someone over the head may simply
be categorized as an aggressive, norm-violating action. However, it becomes
interpreted as a racist act only when the participants in this event are
members of specific groups, and if the aggressor is assumed to act on the
basis of such group membership, for example as sharing specific racist
attitudes.
Against cognitivist or interactionist reduction, these arguments further
emphasize the fact that one cannot escape either the cognitive or the social
dimensions of ideologies, groups and social reality. Both dimensions or
Groups
159
levels are analytically needed to adequately describe and explain the social
'facts' or rather the socio-cognitive constructs of society, groups and their
members.
Inclusion and exclusion
The example of racism shows another important feature of the relations
between ideology and group membership, namely, the social and cognitive
principies and strategies of inclusion and exclusion. Racist ideologies and
practices basically aim at keeping Others down and especially out: out of
our' country, 'our' city, 'our' neighbourhood, 'our' street, 'our' family,
our' jobs and 'our' houses. If some limited forro of admission is accepted,
then only in a lower position: in another (worse) part of town, in worse
housing, worse jobs and so on. Superiority may be denied as the leading
value involved, for example because of official democratic and egalitarian
values. But the implication is always that We, Our Group are self-assigned a
better or higher position and that such a position is deserved and can hence
be justified (Ye were here first', etc.). These ideological principies of
superiority and inferiority may of course be combined with others, such as
those that regulate competition over scarce resources, so that racism typically gets worse in times of an economic recession or other social and
economic pressures on the ingroup.
As we have seen for the discussion of membership, groups thus share
beliefs and practices that regulate inclusion and exclusion. Inclusion may be
made difficult, as is the case for complex initiation rites, or it may be made
easy, as long as the new members identify with the group. Other groups
eagerly go out to recruit new members, as is typically the case for religious
groups and advocacy groups. Overall, we may assume that if groups have
special privileges, that is, preferential access to highly desirable or even
necessary resources (freedom, housing, food, income, employment, etc.)
also the strategies of exclusion will be more forceful. In this case, the stakes
(the interests) being vied for are highest. Typical examples include political
oppression, the forceful exclusion of 'illegar immigrants, or the discrimination of minorities on the labour market. The same is true when the poor are
kept from virtually all resources of society, both the material ones (income,
jobs) as well as the symbolic ones (education, status, respect, culture).
Inclusion and exclusion may also function in a more positive way, for
instance in situations where ingroup solidarity is relevant in the resistance
against domination. Blacks may have black-only organizations in order to
organize against racism, and wornen may have women-only bars in order to
have a place where they need not confront men. Being among 'one' s owñ in
such situations may have a benevolent effect on self-consciousness, the
organization of resistance or simply the reproduction of group beliefs
through conversation. But, as is true for dominant groups and their ideologies, also in this case group formation and identity are closely linked to the
sharing of common beliefs. Exclusion here may foster the development of an
160
Society
ideology that allows group members to evaluate the beliefs and actions of
others, both of the own group, as well as of one or various outgroups.
This informal discussion shows that the notion of group, and the principies and practices of identity and identification, inclusion and exclusion,
access and acceptance, and many other social practices and processes are
intimately linked to fundamental group ideologies. They involve representations of identity, about who We are and what They are, and especially about
what is good for Us and what is not good for Us. Sharing exclusive or
preferential access to scarce resources with others will generally not appear
to be good for Us, unless people can be convinced that marginalization,
discrimination and oppression of Them may eventually also be bad for Us. It
may be bad for business (because good candidates or good business are
excluded), bad for our moral standing (few people want to be called a sexist
or a racist) and eventually bad for our self-esteem if we become convinced
that our values, ideologies, morals or practices are inherently wrong. Group
membership and its ideological basis, after all, are not only about power and
domination, and not only to protect interests, but may also be a source of
pride and pleasure. In the pages that follow, I examine some other features
of these social dimensions of ideologies.
16
Group Relations
Position
Throughout the earlier chapters of this study, ideologies not only appear to
be tied to more or less well-defined groups or movements, but also to
various aspects of relationships between groups. One of the fundamental
categories of the ideological schema therefore also focused on the position
of the group in relation to other groups. Racist ideologies, as we have seen,
are fundamentally based on distinctions being established by ingroups that
simply 'prefer their own' or that feel themselves superior to outgroups, and
manifest themselves in all social forms of problematization, marginalization
or exclusion of the others.
Journalists, as a group, develop professional ideologies primarily in
relation to other elites and other power groups. Thus, they may emphasize
the freedom of the press, oppose censorship, while on the other hand they
will see themselves as the watchdogs of society in the service of the 'public'
at large. Similarly, professors also define themselves as such in relation to
their students, and doctors and lawyers with respect to their patients and
clients. Sometimes these relationships will be more or less egalitarian, but
competitive, in other situations the relationship may be hierarchical and
dominant.
In many cases, various interests of our group may have to be defended or
legitimated against others. And sine conflicts over scarce social resources
may be the very core and function of the development of ideologies, group
position and relations are the most direct social counterpart of ideological
structures, as is most obvious in the well-krtown polarization between
ingroups and outgroups. Indeed, some groups exist by virtue of their
hierarchical or more powerful position, as is the case for superiors and
subordinates, elites and the 'masses' or majorities and minorities. As
discussed in the previous chapter, identification, access and inclusion of
(new) members, may be intimately linked to the exclusion of others, thus
defining power abuse and domination. Let us therefore examine some of
these group relations in somewhat more detall, and see how ideologies are
functionally related to (the reproduction of) these relationships.'
Power and domination
This is not the place to present a new or better theory of power, which has
already filled many studies. 2 In the framework of this chapter I simply take
162
Society
(social) power as a specific type of social relation between groups. Of all the
possible dimensions of this complex notion, I focus on that of control: a
group A has or exercises power over another group B when the members of
A are usually able to control the members of B. This may typically involve
the control of the actions of the other group and its members, in the sense
that the others are not only not (or less) free to do what they want, but may
be brought to act in accordance with the wishes or the interests of the more
powerful group, and against their own best interests (and usually also against
their will). Power relations of age, class, gender, ethnicity, origin,
social position or profession, are typical examples.
Thus, the possession and exercise of (more) power of one group usually
implies the loss or limitation of freedom for the other group. Ideological
claims for freedom, as in freedom of the press, and freedom of the market,
are thus usually claims for power. The same is true, though from a different
perspective, for the claims for freedom — as empowerment — by dominated
groups.
Making others act as one prefers requires resources. Thus, in the most
elementary form of power exercise, namely, that of coercion, the resource
may be bodily (typically male) or institutional (police, military) force. More
sophisticated is the exclusive control over necessary resources (food, housing, jobs or money) by which others may be forced to comply with the
wishes or follow the directives of the powerful. Non-compliance will in that
case lead to undesired consequences (loss of necessary resources), so that
the dominated will have to choose between being dominated but surviving,
on the one hand, or resisting and perishing, on the other. Colonialist and
capitalist oppression and exploitation, as well as traditional socio-economic
oppression, male chauvinism and racism are of this kind.
For my approach to ideology and discourse, a more 'sophisticated' form
of power needs to be dealt with, one that is usually called 'persuasive' and
which is traditionally associated with ideology and hegemony. In this case,
control does not take place (primarily) through physical or socio-economic
coercion, but by more subtle and indirect control of the minds of the
dominated. By controlling the access to public discourse, only specific forms
of knowledge and opinions may be expressed and widely circulated, and
these may persuasively lead to mental models and social representations that
are in the interest of the powerful. Once these mental representations are in
place, the dominated group and its members will tend to act in the interest of
the dominant group 'out of their own free will'. The dominated group may
lack the knowledge or the education to provide alternatives, or it may accept
that the dominance of the dominant roup is natural or inevitable, and
resistance pointless or even unthinkable.
In this study it is this type of discursive and ideological control that will
be taken as the main example of power and dominance, one that seems
prevalent in contemporary'information and communicatioñ societies, in
which knowledge and the access to the media and public discourse are the
crucial resources to control the minds, and hence indirectly the actions, of
Group relations
163
others. It is here that consent and consensus play a fundamental role in the
exercise of power and the reproduction of ideologies that support such
power. Obviously, those who have persuasive, ideological or discursive
power, also usually have the coercive powers to take care of those who
woñ t comply with the directions of symbolic power. Economic and
physical means may then be applied where less blatant power means fail.
Although the notions of power and domination seem to be used as
synonyms aboye, I use them in a different sense. Since domination implies
involuntary inequality, I reserve it as a shorthand for abuse of power. This
also implies that I doñ t use power only in a negative sense: power may be
consensual and beneficial, as is the case in situations where groups elect
their leaders and temporarily accord special power to them. Domination,
then, presupposes power and deviance from general or universal ethical
principies, that defines abuse, for instance the exercise of social power in
oné s personal interest, hurting other people, and so on. Both power and
domination, as relations between groups, need to be based on ideologies in
order for such relations to be reproduced in everyday life and the mundane
practices of group members.
This may of course involve all kinds of variation, gradual differences
between power and counter-power, and the more or less harsh or soft
exercise of power, or the more or less tough resistance or compliance by the
dominated. It is in this more contextualized way that power is sometimes
said to be 'everywheré There would be no dominant groups if power were
not exercised, sometimes very subtly, through everyday practices. Moreover,
there are (members of) dominated groups who comply, and dissident
dominant group members who show solidarity with the underdog. Despite
these variations and the uneven exercise or distribution of domination and
resistance, we may assume that, at a higher level of analysis, relations of
domination between whole groups exist, and that ideologies control these
relations and their everyday implementation.
Within this framework, then, we first need to examine the role of
ideologies in the reproduction of power and dominance. Indeed, one of the
core notions of classical ideology analysis and critique has always been that
ideologies are developed and applied as legitimation for the abuse of power
(domination) and its resulting social inequality.
In my analysis I have started from the assumption that ideologies are
systems of basic principies that are socially shared by groups. Such
ideologies have a number of cognitive and social functions, including the
maintenance of group cohesion and solidarity, as well as the protection (or
acquisition) of scarce social resources. In sum, socially, ideologies are
developed in order to make sure that group members think, believe and act
in such a way that their actions are in the interests of themselves and the
group as a whole. Such a 'co-ordinative' social function is in the interest of
the group in its relationships with other groups.
If a group is in a dominant relationship with respect to other groups, for
instance on account of its privileged access to social resources, ideologies
. 4
164
Society
have the double function of maintaining or confirming the status quo, and at
the same time of providing the basic cognitive framework for argumenta to
persuade its own members as well as others that this situation is Juñ ,
natural', God-given, or otherwise legitimate.
Thus, priority in employment and housing for 'our own' people may thus
be legitimated by the racist principie of ethnic or racial superiority, by the
commonsensé nationalist principie that 'our people' should of course have
priority over newcomers, or by the opportunistic socio-economic reason that
there is a shortage of houses and jobs, and that 'objective' criteria must be
applied for 'fair' decisions, and that those who come last have fewer rights
than those who were already
Thus, we see how power and domination, as a specific form of inter-group
relation and societal structure, may be reproduced by various ideologies (at
the socio-cognitive level) and by the social practices (at the microsocial level
of situations) that'implement' such ideologies. Whether these social practices already existed before they were legitimated by an ideology, or whether
they only can be thus organized because of an ideology, may be a moot point
in practice, asking the proverbial chicken and egg question. Rather, we
would say that the dynamics of the interplay of cognition and social practice
shows that they mutually constitute each other in a 'dialectical' process.
Here power abuse is sometimes ideologically justified afterwards, but at the
same time (socially or historically acquired) negative attitudes against others
may already exist in order for power abuse to be exercised in the first
place.
The primacy of ideology over action
Theoretically and historically, the question of the primacy of ideology over
action (or vice versa) is less frivolous. It has for instance been asked in
relation to the system of slavery, and its abolition: were racist ideologies
(e.g. about the attributed inferiority of Africans) invented to legitimate
slavery and colonialism, or could Africans be enslaved in the first place only
because they were already seen as inferior to Europeans?
Although this is not the place to answer such questions, a socio-cognitive
theory of ideology would opt for the latter suggestion — enslavement
presupposes knowledge and opinions about peoples that may be legitimately
(ethically, etc.) enslaved: for instance non-Christians, people from a different continent or country, people with a different appearance, or simply
people that were conquered by 'lis', as the history of slavery (also of others
than Africans) has shown. These criteria of difference were generally
associated with negative opinions about the others, or at least with feelings
of superiority of the own group. Hence, engaging in enslavement already
presupposes some kind of negative attitude about the outgroup, which
allowed slave-traders and slave-owners to legitimately do what they did, for
example without being sanctioned by the state or the Church. If not, they
Group relations
165
could simply — and probably more cheaply — have enslaved people of their
own group, namely, the poor, as was the system of capitalistic exploitation
that followed the slave system or as happened with indentured whites.
However, precisely because of increasing ethical (and at the same time
economic) argumenta against slavery, it became necessary to farther develop
the ideological system that legitimated slavery. Various pseudo-scientific
reasons, for example, about the differences between the 'mes' were
adduced as foundations for such ideologies, thus giving rise to more specific
and explicit racist ideologies, where earlier, at least until the eighteenth
century, the inferiority and hence 'enslavability' of the others was simply
taken for granted — and hence ideologically presupposed. 5
My point here is merely that systems of social practices of groups (and not
incidental actions of individuals) tend to be oriented towards the interests of
these groups, and such a co-ordination problem can only be solved if the
group shares specific knowledge, attitudes, norms, values and ideologies in
the first place. These may be very simple and elementary in the beginning,
but without them social practices would be more or less haphazard and
individual. Concerted action in favour of the group and at the same time for
its members, thus, primarily presupposes shared social cognitions, and not
the other way around.
Legitirnation based on such ideologies only becomes relevant when
needed, that is, in contexts of opposition, critique and social struggle. They
are social (discursive) social practices in their own right, and their absence
does not imply absence of ideology, but only that in such a case the ideology
may simply have been taken for granted.
Pure power abuse, thus, does not always need social (discursive) practices
of legitirnation, but it always does need belief systems in order to coordinate the social practices that keep the system of domination intact. In the
case of slavery and exploitation, thus, negative actitudes and ideologies
about relevant outgroups are needed to subject outgroup members to the
social practices of domination. As is the case with most complex social
actions of groups, ideologies are also necessary as fundamental guidelines
for the management of domination.
Of course, once systems of power and domination are already existing, the
relationships between social practices, social relations of domination and
inequality on the one hand, and attitudes, norms, values and ideologies on
the other hand, will mutually sustain each other. Thus, slavery was abolished
precisely for this double-edged reason: it did not pay (enough) anymore,
while at the same time the ideological justification was successfully challenged by abolitionists and their supporters. In such complex social situations, causes and consequences, actions and minds, are difficult to keep
separated. Yet, for purely 'psycho-logical' reasons, I assume that people
cannot act rationally and purposefully withoút the appropriate social cognitions. At the level of the maintenance of groups, group interests and group
relations, such cognitive conditions require the development of attitudes and
ideologies.
166
Society
These ideologies may themselves be sustained by (successful) social
practices, but they are not only 'invented' as a consequence of such actions,
for example as forms of post hoc justification. They may be acquired simply
by discourse, communication or perception, much in the same way as the
Europeans of over five hundred years ago 'knew' about Africans through
tales, myths, histories, travelogues, and later through 'scientific' discourse. It
is that complex, but essentially 'biased' — and later constantly updated —
image that was at the basis of the social practices that led to the slave
system, even if these were not the only social cognitions that informed such
decisions: Of course also socio-economic, geographic, and other belief
systems and conditions were involved in these decisions. Many other
examples of dominance systems in society and their historical growth,
change and demise may thus be explained also as a consequence (ánd not the
cause) of developing or changing ideologies.
As I have shown aboye, even 'objective' socio-economic circumstances,
as such, do not influence social actions directly, but only through their
(mental) interpretation and representation. Thus, there are most certainly
also powerful social and economic conditions that allowed or favoured the
growing feminist movement of the 1960s and 1970s, but it seems historically more correct to maintain that the major'causes' of that movement
were ideological, and brought about by politicians, writers, academics,
artists and other women (and some men) who advocated equal rights for
women. This happened in a period in which also other forms of ideological
change took place, such as the civil rights movement, decolonialization, and
challenges to the authoritarian state.
This suggests that the relationships between power, dominance and
ideologies need to be analysed carefully, and 1 already assumed that
ideologies may not always (or even seldom) be Invented' post hoc to
legitimate patterns of inequality and the social practices that constitute such
inequality. Legitimation is usually discursive and often argumentative, and
we saw that it may be especially required in specific social contexts, for
example of opposition and struggle. However, such opposition itself logically follows the existence of domination, and domination is possible only
with at least a minimum of shared social cognition, and hence by ideologies
of dominant groups about dominated groups. That ideologies may change as
a result of such opposition, and indeed as a consequence of the ideological
debate that accompanies such resistance, is obvious, but again suggests that
ideologies are more or less autonomous, and may be changing as a
consequence of other ideologies and their manifestations in public discourse,
and not (always) as a consequence of changing social practices.
Indeed, traditional systems of power were usually coercive, that is, based
in physical action control, violence, military power, or the practices of the
secret police or strongmen. On the other hand much 'modem' power is
persuasive, discursive and (hence) ideological. Dominant groups no longer
maintain their position only by force or even threats of force (the latter
already being forms of discourse), but by complex systems of discourse and
Group relations
167
ideologies that make (most members of) dominated groups believe or accept
that domination is justified (as in democratic systems), natural (as in gender
and racial domination) or inevitable (as is the case for the socio-economic
grounds and the logic' of the market).
As soon as some and especially many members of dominated groups no
longer accept such ideological grounds, and have acquired the symbolic
means to propagate counter-ideologies and the material conditions to act
upon such counter-ideologies, ideological change will be inevitable, and
changes in social practices will (sometimes very slowly) follow. Indeed,
many men will today accept at least some basic tenets of feminist ideologies
according to which women and men are equal and should be treated equally,
but it is well known that their social practices do not yet always meet the
precepts of this new gender ideology. That men are aware of such changing
ideologies is frequently apparent from their discourses, for example in
disclaimers such as 'We do not discriminate against women, but.. or
have tried to find a woman, but. .'. That is, disclaimers of this nature, to
which we shall be coming back in the next part of this study, are typical
expressions of the contradictions, if not the moral dilemmas, between
official or dominant ideologies and actual practices, talk and text. At the
same time, the disclaimers obviously function as moves in face-keeping
strategies of positive self-presentation.
In sum, despite the complexities of the (sometimes mutual) relations
between ideologies, power and domination, the theoretical framework
assumes that, historically and theoretically, ideas precede actions, and (at
least simple) ideologies precede the systems of social practices that define
domination. But, once the system of domination is in place, and especially
when it is being challenged, then ideologies may well be further developed
to provide for the legitimation of the system. This does not imply, however,
that ideologies only serve as systems for discursive legitimation, which
would suggest a post hoc role of ideologies. More relevantly, ideologies
monitor and organize group knowledge and attitudes and hence the beliefs
that members need in order to construct the models controlling the actions
that implement domination.
Practices of power abuse, domination and oppression can be effective
only when co-ordinated, when relevant model structures are socially shared
— and ideologies precisely serve that 'practical' goal. As soon as ingroup
members need to be recruited and persuaded to share in the actions, against
outgroup members, which they would not undertake against ingroup members (which by itself presupposes social norms and attitudes about what
should or should not be done), these underlying ideologies may need to be
discursively expressed and detailed even for 'internar use and the intragroup reproduction of power and dominance.
Dominance thus requires a fair amount of consensus as well as practical
co-ordination, and ideologies are needed both for the maintenance of the
relations of power with respect to the others, as well as for the maintenance
of ingroup representations that allow such consensus to be reproduced in
168
Society
everyday life — and to marginalize or punish deviants and dissidents that
may threaten, as the 'enemy-within', the dominance of the ingroup. The
anti-communist scare of Joe McCarthy was precisely designed to protect and
maintain the anti-communist consensus and coherence of a country that
represented itself as besieged by World Communism.
This suggests that pattems of power and domination and their underlying
ideologies also apply within the group itself, namely, between the elites and
the rest, between the leaders and the led, between the thinkers and the doers,
a point that needs to be discussed separately later. This will also allow us to
reflect about who actually 'invents' the ideologies shared by groups, and
whether ideologies are spontaneous popular constructions, or rather those of
ideologues or intellectuals who conceive of them first.
Another point to be dealt with (again) in this framework is the wellknown question whether ideologies are essentially associated with domination and dominant groups, or whether we need a more general notion of
ideology for any kind of social group in a specific social position, including
that of resistance.
Conflict and struggle
Domination usually leads to resistance and struggle to overcome inequality
and oppression. It is common practice in the study of ideology to associate
ideologies with domination and its legitimation. I proposed that also
resistance needs a socio-cognitive basis in terms of group-relevant values,
principies, ideologies and its more specific knowledge and attitudes. In the
same way as the exercise and co-ordination of power abuse needs an
ideológical basis, also group-intemal solidarity and inter-group resistance
needs to be ideologically organized. Whereas it may be in the interests of a
dominant group to conceal their power abuse and to hide the forms of
quafity that are its consequences, dissidents and opponents may be specifically interested in uncovering and exposing domination and inequality, and
to manifest and legitímate as 'juse their own, counter, ideologies. Indeed,
that was the point of the communist 'manifestó , as it was for many other
manifestos and declarations (like that of the various declarations of human
rights) in the first place.
From a critical point of view, this may well imply that dominant groups
favour falsehood, deceit and manipulation, and that dominated groups
advocate truth, openness and rational or emotional persuasion, that is, goals
with which also scholars may want to show agreement. Since also most
scholars define themselves (ideologically) as people who want to describe
objectively' the real social relations involved, their interest may in this
respect sometimes be consistent with the subjective, self-serving truths of
oppositional groups. However, since their ideologies of class and profession
may at the same time be inconsistent with the interests and the demands of
the poor, the left, the women, or the minorities, most (middle-class, white,
1
Group relations
169
male, etc.) scholars at the same time prefer to ignore such demands and to
strategically look elsewhere and do their 'objective' research on lessthreatening topics.
Hence the insistente on (scientific) truth in much oppositional ideologies
and critical studies of ideology. We also know, however, that in many social,
economic, political and ideological conflicts, the distinction between truth
and falsity is not that clearcut. This and many other theoretical reasons
suggest that it is more adequate to adopt a general concept of ideology, and
to assume that ideologies by definition represent the interests of a specific
social group, whether or not (in our view as observers, critics or participants)
the group's beliefs are based on true social analysis, justified claims or
legitimate action.
If ideologies represent group interests, and if conflicting interests also
imply social conflict of various kinds, it seems logical to assume that
ideologies by definition imply conflict. For fundamental group relations such
as those of class, gender and ethnicity, this seems hardly controversial: the
empirical facts of the international class struggle, the womeñ s movement
and the civil rights movements, hardly allow an other conclusion. Conflicts
of interests here are so fundamental that open social conflict is a matter of
everyday life, and much of this conflict is not only about socio-economic
interests, but also about symbolic, ideological ones.
But in the same way as I asked before whether all social groups have
ideologies, I should now ask whether all social conflicts between groups are
ideological, and whether all ideological differences always lead to social
conflict. Theoretically, groups may have different and even conflicting
ideologies, but have learned to live with these in relative social peace.
Indeed, there may be higher-order goals and interests that prevent social
conflict between two groups. This is not merely a question of principle, but
also an empirical matter.
Thus, whereas in some societies or cultures, religious differences may be
the basis of acrimonious, open conflict (as in Northern Ireland or India), in
others mutual religious tolerance may be prevalent. Similar examples may
be given about linguistic or other cultural conflicts. Of course, such an
empirical question may hinge on the definition of the very notion of conflict.
If conflict also includes mere differences of opinion and debate, then
virtually all ideological differences will be conflictual. However, if we limit
conflicts to any form of dominante, to one-sided or mutual discrimination or
other social practices in which ingroup members are favoured over outgroup
members in social interaction, then we have a more specific notion of
conflict that may be relevant for a more selective use of the combination of
ideology and conflict.
It is in this more restricted sense, then, that we might maintain that
ideological differences do not necessarily lead to open social conflict.
Professors and students, doctors and patients, lawyers and clients, different
political groups or parties, non-governmental organizations and action
groups may all have different and inconsistent or even conflicting interests
170
Society
and ideologies without therefore exhibiting such conflict in forms of
discriminatory or oppressive practices directed against outgroup members.
In other words, whereas most social conflicts and struggle presuppose
ideological conflicts (especially over scarce resources), the opposite is not
trae — not all ideological conflicts imply struggle and social conflict.
Ideologies may incite to self-serving group actions, but laves, norms,
agreements or other, non-ideological self-interests, may prohibit open conflict. Sometimes social peace and co-operation may be the prevailing, also
self-serving criterion over sectarian or ideologically based open conflict. In
that case, the ideological struggle may be transferred to the level of mutual
discursive persuasion, negotiation and consensus policies.
Competition
Indeed, one form of ideological conflict that need not imply social conflict
may be based on inter-group competition. Different groups may have the
same goal, but want to realize it with different means. Peace, equality,
human rights, the equal distribution of wealth, and so on, may be ultimate
goals that countless groups and movements, with different ideologies, may
want to achieve in different ways. Such groups, trying to realize the same
goals, or vying for the same social resources, may just be competitive and
not be in open conflict with each other. Indeed, this is the very ideal
(idealistic and ideological) principle of liberal market philosophies.
The question may then be raised again: does social competition require
ideological foundations, given the differences of goals or interests, and vice
versa, do all ideological differences at least imply some forro of competition? I think the first question should be answered negatively. First, because
competition is not necessarily social and group-based, but may also be interpersonal, and, second, because competition may also exist between groups
with the same ideology, as would typically be the case for different
companies in the same social domain vying for the same customers.
Differences here need not be 'deeply' ideological, but rather practical and
strategic, that is, different ways of reaching the same goal and following the
same principies.
On the other hand, competition between different political parties during
an election, or between two different ecological groups, may well be based
on ideological conflicts. This suggests that the second question may well be
answered positively: ideological differences between groups usually imply
competition, if only when vying for membership and the recruitment of new
members, or the persuasion of outsiders. More common is of course the
competition for scarce social resources, such as residence, income, housing
and welfare on the other hand, and non-material resources such as knowledge, education, esteem and status, on the other. Thus, struggle and open
conflict, while based on conflicting interests, usually implies competition,
but not vice versa.
1
Group relations
171
Co-operation
We may make a foral theoretical step and ask whether also inter-group
relations of co-operation may be ideologically based. This certainly seems to
be the case. Two groups or organizations may have different ideologies (e.g.
Catholics and Muslims), but may well co-operate to realize a common goal,
and jointly acquire or defend shared interests (e.g. support for religious
activities and freedoms, or the prohibition of abortion). Ideological opponents may thus become allies in pursuing the realization of the same goals.
But whereas open conflict and struggle may need ideological foundation as
such, especially in categorizing the beliefs about own group position and the
relations with other groups, co-operation as such does not need ideological
support. One common goal or one important attitude or opinion may be
enough to organize the joint accomplishment of social action.
Conclusion
From this discussion it may be concluded that inter-group relations are
generally fundamental in the development and support of ideologies, and
conversely that ideologies are at the basis of the social practices that
implement such group relations. Conflicts of class,'racé and gender thus
pitch dominant groups against (usually) minority groups or groups with less
power. These conflicts are usually about access and control over material or
symbolic resources. Other conflicts, as well as competition and co-operation
between groups, exist but do not seem to be ideological, but rather practical,
for example when groups engage in different ways to separately or jointly
realize a common or a related goal. Conversely, although ideologies often
imply struggle and conflict, this implication does not always hold: ideologies
that are in conflict do not necessarily lead to, or emerge from, social struggle
and conflict, but may also be needed to manage diversity.
Who 'invents' ideologies?
In order to complete the picture of the social basis and dimensions of
ideologies, we should now ask where ideologies 'come from' in the first
place. Who, indeed, Invents' ideologies? Or do they arise and develop
spontaneously in a group, as a form of jointly produced social cognition that
has no specific authorship, as would be the case for a natural language?
Many ideologies seem to emerge from large groups of people, if not from
the 'masses'. Ecologist, feminist, socialist, nationalist or capitalist ideologies
are examples of ideologies that are shared and carried by many people, often
across national boundaries and continente. That these should be 'invented'
by specific individuals, or by a small group of Ideologues', thus seems to be
counter to the basic conception of ideologies as shared, social belief
systems.
One question, often formulated in political psychology, is that it remains
to be seen whether such large groups of people do indeed have a more or
less explicit or articulated ideology in the first place. They may share a few
general principles and goals, but not a 'complete' ideology. Such more
detailed and explicit ideologies are then typically attributed to the leaders,
the intellectuals, the elites or indeed the 'ideologues' of such groups.'
As is the case for social and personal differences of knowledge, we may
expect variations of attitudes and ideologies within the same group. Experts
have access to more and more varied forms of discourse, 2 may communicate
more often and more explicitly about the ideologies of their group, and may
therefore develop more detailed and more 'articulate' ideological systems.
They may be more familiar with ideological arguments against their ideological opinions, and may therefore become more proficient in ideological
counter-arguments, which again may contribute to more detailed attitudes
and ideologies. In other words, explicit ideological practices as well as
ideological discourses are systematically related to ideologies, which mutually may facilitate each other. Leaders, intellectuals and other'ideologues'
of a group typically may be expected to play such roles, especially because
of their privileged access to public discourse, and because of their tasks to
lead a group, co-ordinate its actions, and make sure that its goals are realized
and its interests protected.
At the same time, there is no clear-cut distinction between such 'ideologues' and the other members of a group. Any member who is more or less
Elites
173
conscious of her or his group membership and its goals, and who is able to
participate, even passively, in public ideological discourse (e.g. by reading
editorials in the press), may thus become fairly proficient in the argumentative expression of underlying ideologies, and thus develop detailed ideologies. Thus, in the women's movement, not only the leaders, intellectuals,
experts or other 'ideologues' may develop ideologies, but also other relatively active and 'conscious' members. After all, if ideologies are constitutive of %ved experiences' and common sense, most members will be
confronted with ideological practices, and may in principie be able to
interpret these actordingly. 4
There are probably differences between different ideological groups in
this respect. Political party members may be less ideologically conscious
about their party membership than are members of religious groups or social
movements. As a criterion for such differences, we may assume that the
nature of socialization in the group, the amount of top-down or mutual
indoctrination, the number of meetings and other forros of active participation, as well as the nature of everyday experiences that have an ideological
basis, will all contribute to more or less explicit ideologies. Opposition
groups, and social movements who have access to public discourse may thus
quickly raise broad support for their ideological beliefs, and thus make
members more 'conscious' about the reasons for their group membership.
The women's movement, the civil rights movement and the ecológical
movement from the 1960s to the 1990s are typical examples in point.
Moreover, I do not conceive of ideologies only in tercos of explicit,
detailed systems, for example those of the Ideologues' of a group. A few
basic principies that organize the attitudes of group members may be enough
to define a core ideology, which in turra will influence social practices and
discourses. Thus, a fundamental value like 'equality' applied to gender
relations, will yield the basic ideological proposition, Women and men are
equal.' Such a proposition may be enough for more specific application in
attitudes about equal rights in general, for example in voting, employment,
promotion, salaries, family roles and a host of other social practices and
situations. In other words, no very sophisticated, theoretical analysis is
necessary in the 'invention' as well as the application of ideologies.
Sometimes, a single basic value such as equality or freedom may be enough
to construct an ideology when it is applied to the evaluation of the position
of the own group.
What is crucial, though, is access to public discourse. For some social
movements, such discourse may literally begin with shouted slogans in the
streets. But, in general, groups and social movements historically have their
basis in the writings of smaller elite groups of philosophers, writers,
academics, politicians, union leaders and other elites who have at least some
access to books or the mass media. These writings may be based on critical
social analysis, values and other ethical principies, as well as on personal
experiences shared with other members of the group. Whereas the latter case
is typical for the women's movement and the civil rights movement, group
174
Society
membership of leading elites is not essential, as long as these elites are able
to express and articulate the goals, interests and, vicariously, even the daily
experiences of the group 'for' which they write and take action. This is
typically the case for the class struggle, but also for anti-racists, or people in
the North who feel solidarity with those oppressed in the South.
Top-down or bottom-up?
Related to the question whether ideologies are rather the systems known by
Ideologues' or other elites, or also (fully) shared, as such, by the population
at large, is the question of development and influence. That is, it is not
merely relevant to investigate where ideological beliefs come from, but also
how they are shared and communicated.
We have seen that historical evidence suggests that at least many
ideologies first seem to be invented and propagated top-down: a small
number of more or less conscious and articulate leaders, intellectuals, or
Ideologues' tend to formulate the ideological principies of a group. Then,
through various forms of intra-group discourse (debate, meetings, propaganda, publications) and other institutional and organizational practices,
such ideologies are slowly propagated among group members and society at
large. Indeed, as we have seen, only the leaders or other elites may have
access to the means of communication and public discourse that allow
propagation and the reproduction of ideologies in the first place.
This assumption probably applies to such broad social movements as
liberalism, socialism, feminism, and the environmental movement, among
others. Sometimes rather precise historical and even personal antecedents of
ideologies may be found, for instance with eighteenth-century French
philosophers, or twentieth-century African-American leaders. Specific books
of specific authors may spawn a movement and its ideological grounding.
Although ah this may be true, it seems at the same time inconsistent with
the social, group-based nature of ideologies. If ideologies are inherently
social, how can they be 'invented' by individuals? This would historically
reduce social movements and their struggles to personalistic initiatives,
actions and ideas.
My view of this dilemma is that there is no contradiction here. Specific
ideas may well have been 'invented' by one or a few individual thinkers,
revolutionaries, writers or other elites. But for such a set of 'ideas' to
become an ideology, in my definition, it must essentially be socially shared.
One major condition for this process of social sharing and reproduction is, at
least in general, that the group members are able to identify with the group
and its ideology. Its goals, practices, position, values, and so on must also
apply to them, and be relevant for their everyday experiences. Socialist or
communist ideologies thus applied to the everyday lives of workers, and so
did feminist ideologies for the everyday lives of women. In other words,
even when 'ideas' or arguments for such ideas may initially be 'invented' or
Elites
175
at least voiced publicly by specific individuals, they may constitute an
ideology only when shared and 'carried' by a group of people whose
interests are related to there ideas in the first place.
This also suggests that the success and acceptance of some basic opinions
as an ideology by a group may indeed presuppose relevant experiences of
group members. Gender inequality and oppression already existed before the
womeñ s movement, and at least some women were conscious of such
relationships and resented them. Feminist ideas about equality and autonomy, partly borrowed from similar ideas in the realm of politics, were thus
hardly more than the explicit formulation of more or less implicit ideas
about 'what was wrong' and 'what had to be done' among many wornen. In
that respect, the leaders of the womeñ s movement were inspired both by
their own experiences as wornen and by (initially maybe anecdotal) information about and direct observations of the experiences of other women. It is in
this respect that elite ideas and the invention of ideologies are not merely
conditioned by the acceptance of such ideas by social groups, but at the
same time by the very experiences and (possibly informal) discourses of
group members in the first place.
That is, initial explicit and public formulations may have taken place
especially by a few leaders, elites or intellectuals, but the opinions, attitudes
and experiences on which they are based may already have been widely
shared by larger groups, and may already have given rise to occasional,
isolated forros of protest, resistance or dissidence among such larger groups.
In this respect, the development of ideologies is indeed a social, two-way
process, in which top-down leadership and influence is closely tied to
bottorn-up influence, experience and action.
Elite discourse that does not formulate popular opinions is not likely to
spawn a popular movement. And once such a movement grows there are
numerous ways 'ordinary' members are able to make themselves heard to
the elites in mass meetings, demonstrations and other forros of public action.
Indeed, those leaders and elites will generally be most influential who are
best able to articulate the concems and the experiences of the group as a
whole. And conversely, the historical record also shows that grassroots
experiences and opinions alone may not be, a sufficient condition for the
articulation of explicit ideologies in public discourses that are able to
influence wider social debate and to lead to social change, for example
among those groups (and their leaders) who initially oppose a popular
movement.
This top-down, elite influence is especially noteworthy for those cases
where interests and everyday experiences are initially found to be less acute
and fundamental for a large group of people. This is for instante the case for
the environmental movement, in which the public at large initially was
barely aware of the conditions and consequences of pollution. Only when
the threat to health and survival, both to that of humanity as well as to that
of nature, was clearly demonstrated by research (the Club of Rome) and
concrete examples (like Chernobyl), could environmentalism become a
176
Society
popular movement. Awareness and consciousness, thus, often is a social
construct, and elites may play a role in the invention of such a construct.
Elite racism
Racism is a well-known example of the complexities of the relations
between elite ideas and popular resentment. Research shows that white elites
emphatically deny their role in the reproduction of racism, while at the same
time blaming poor whites for xenophobic resentment and taking advantage
of such resentment to propagate their own ethnocentric or anti-foreigner
ideas and
i 6 Racism thus is also essentially top-down and bottom-up.
Bottom-up influences are generated by the everyday socio-economic experiences of poverty, run-down inner cities and unemployment, and the (biased)
perception of 'easy' immigration and the alleged favouring of immigrants in
employment, housing and welfare. In other words, popular racism and its
ideologies are primarily based on the perception of unfair competition for
scarce material and symbolic resources.
However, this is only part of the story of racism. Research also shows that
xenophobic beliefs are not always or not only limited to poor whites in a
social—economic predicament. Indeed, prejudices and discrimination may
even be more widespread not so much at the bottom of the social hierarchy,
but just one or two rungs higher, for example in the lower middle class, as
relative deprivation theories would predict. Here, instead of feelings of
competition, the fear of loss of barely acquired resources may be stronger
than among those at the bottom who have nothing to lose.
But even this common observation only provides one more element of the
complex structure of racism and its ideological basis. Indeed, prejudice and
discrimination, though of different types, are widespread throughout white
society, also among the elites themselves. Whereas the confrontation with
other peoples, languages and cultures as such may be much more familiar
among (travelling and reading) elites, this does not mean full acceptance in
everyday life of 'racially or culturally different others, for instance as
colleagues or bosses. That is, the general social superiority feeling of class
or education among the elites easily transfers to those of race and ethnicity.
Instead of the competitive 'threat' to jobs or housing, elite racism is thus
much more oriented towards cultural issues, such as habits, religion,
language, education and values. The world-wide construction of the threat of
Islam, for instance, is not a popular movement, but an elite phenomenon.
Widespread discrimination on the job is also managed by elites, namely, the
managers. Bias, stereotyping and outright ethnic polarization in the media is
the product of journalists, or of the politicians they use as reliable sources,
and hence also an elite phenomenon. The same is true for biased textbooks
and scholarly research.
In sum, wherever it really counts (immigration, residence, housing, jobs,
education, media, health care, welfare, or the arts) the crucial decisions
Elites
177
about inclusion and exclusion are made by the elites. It is therefore
essentially the elites who pre-formulate many of the everyday ideological
beliefs that have become widespread in racist societies. These ideas need not
be explicitly racist, unlike those of extremist right-wing scholars who
legitimate ethnic inequality through pseudo-research. Although such scholars are often marginal, they may have tremendous influence on racist
organizations as providers of scholarly legitimation.
Rather, I am referring to much more mundane beliefs and arguments
against immigration and the multicultural society, beliefs that are easily
accepted by everyday common sense, even of those among the population at
large who have no daily dealing with minorities. To blame immigrants,
refugees and minorities for generally felt problems of, for example, unemployment, overpopulation, inner-city decay and the destruction of the
welfare state, is relatively easy as long as the mass media and many
intellectuals comply, at least in a moderate way. Having generated the
popular resentment against foreigners that may follow such subtly racist
propaganda, this popular resentment may be used again as a 'democratic'
legitimation against immigration, equal rights or affirmative action (see
Chapter 28 for a concrete illustration of these and related strategies).
Given the virtually exclusive access to, and control over the mass media
by the elites, and the marginal role of ethnic minorities and their economic
competition in the everyday lives of most white people, it is difficult to
accept that white racism is a spontaneous, popular movement. Indeed, if the
elites were consistently and fundamentally opposed to any form of prejudice,
stereotyping and discrimination, ah the decisions that really count for
minorities would not be so consistently against them, as immigration
restrictions, job discrimination and bias in reporting and textbooks shows.
Indeed, if only and originally popular in origin, racist beliefs would not have
access to anti-racist public means of discourse in the first place.
From these arguments (and much research) it follows that whereas racism
may at first sight appear as a form of popular resentment, with only small
intervention of some marginal, prejudiced elites, or even as a forro of
inequality that permeates the whole of Western societies, in fact it is largely
based on elite ideologies, discourse and social practices. Ideologies preformulated by these elites, however, may in the appropriate socio-economic
circumstances be incorporated in initially vague and undirected popular
resentment. Such resentment and its socio-economic basis may be such that
the elites (and especially the politicians) in tum may often be seen as'soft'
on immigration or minorities, on the basis of a more moderate style of public
discourse, as propagated through the media.'
That is, one might say that in ethnic matters, large segments of the
population are able to read between these 'moderate' fines, and expect the
more blatant forms of anti-immigrant beliefs as they are voiced in much
everyday talk in private situations. Hence, there is no contradiction here
between strong popular resentment and moderate elite discourse about
immigration and minorities. On the contrary, what is being presupposed or
178
Society
implied by the elites, as well as the actual social practices of discrimination
or exclusion by these elites, is enough as a legitimation for popular
resentment in the first place. If leading politicians or newspapers focus on
the many problems of immigration, and advocate various forms of exclusion, then many people will feel vindicated in their resentment against 'those
foreigners' who are blamed for fundamental social and cultural problems
and insecurity.
Another proof of such top-down, elite influence in the reproduction of
racist ideologies and practices is that in those situations where leaders do
take energetic anti-racist positions, also their subordinates or group members
tend to follow and accept such beliefs and policies. Although this may not be
the case for all social issues, and although ideological influence may
sometimes be both top-down and bottom-up, racism seems to be a rather
clear case of predominant elite influence. One other reason for this special
case is that ethnic prejudices and ideologies have to do with fundamental
values of equality and social and cultural acceptance, and less with economic threats and the experiences of everyday life. Minorities are literally
of minor consequence in most Western societies, and the consequences of
inter-group relations in this case are therefore of a symbolic and ideological
rather than of a socio-economic nature. The point is precisely that the elites
transform socio-cultural interests into socio-economic interests that may be
acceptable to the population at large, for example by blaming social
problems (like unemployment or crime) or a lagging economy on immigrants.
Acceptance, tolerance and diversity (and their counterparts) are typical
elite issues, and as long as the elites do not wholeheartedly accept the
multiculturalization of white Western societies, it is hardly likely that such
will be the case among the population at large. The massive presence of
ambiguous or negative news, movies, advertising, opinion, social debate or
political propaganda about minorities and immigration — all managed by the
elites — in this case finds an easy target among those of the population who
only too readily accept that prejudice and discrimination of the others are
only in their own best interests. Indeed, racist ideologies are so easy to
produce and reproduce precisely because of the elite control of the mass
media who specialize in the communication of largely symbolic ideologies,
and because racism as a system of inequality is in the interest of all (white)
group members.
Dominant Ideologies?
Introduction
A major debate in the study of ideologies pertains to the question whether
ideologies are by defmition 'dominant', or should be defined in broader
terms, independent of whether or not groups are dominant, or whether
ideologies are able to 'dominate' the minds of all people in the first place. In
the previous chapters it has already been suggested several times that a
general theory of ideology should not limit the notion to dominant ideologies. However, this decision needs to be discussed in somewhat more
detall in this chapter.'
Following the dictum of Marx and Engels about ruling ideologies being
the ideas of the ruling class, it is frequently debated whether such 'dominant'
ideologies exist in the first place, whether the dominant 'class' has a unified
ideology, and whether or not such ideologies are able to control the
(ideologies of the) dominated classes. Similar questions may of course be
formulated for other relations of dominance, that is, also for gender,
ethnicity, and so on.
Several of the notions earlier discussed with respect to the top-down
influence of ideologies and the role of the elites are combined in these
questions. This also suggests that, as such, these questions may well be too
general and too broad, and may be answered only in a more analytical
fashion.
Groups of various kinds (here including classes) develop group ideologies, and do so especially in social structures characterized by conflict,
competition and dominance. On a very global level,' nothing seems more
obvious, then, that if there are 'dominant classes', such classes also will have
their own ideologies. The questions which then need to be asked first are
what these classes are and whether, how and whom they dominate. Thus, if
the 'rich' are such a class, we may assume that they will develop an ideology
that is geared towards the maintenance of their special access to social
resources, such as capital, income, tax breaks, status, and so on.
However, if (at least leading) politicians, corporate managers, scholars,
joumalists, professionals and other elites are also part of 'the' dominant
class, or form their own dominant classes, then we have a complication. Will
they develop an overall ideology shared by these groups or 'classes', or will
they each tend to develop their own, more specific ideologies, tailored to
their own interests, position, goals and power?
180
Society
There is no reason why both should not be the case. Obviously (leading)
journalists, scholars and politicians have different interests, and will therefore (also) develop specific, group-based ideologies, as discussed aboye.
However, they may well all have a number of interests in common, for
example as related to their (usually middle-class) position and power. Such
partly shared ideology fragments may for instance pertain to their specific
access to scarce resources (income, employment, housing, status, knowledge, power), identity and membership (as elites or leaders), and especially
their position relative to non-dominant groups (variously defined by them as
the 'masses', the 'public', the 'voters', 'ordinary people', and so on).
That is, despite possíble competition and conflicting interests, some
ideology fragments may be shared in a common, overarching, 'dominant'
ideology. Whether or not such shared fragments exist at any one time and
social situation, is an empirical matter, but it seems quite plausible that if
dominant groups have at least their 'dominance' in common, they will also
have the corresponding ideology fragments in common that sustain and help
legitimate such dominance Indeed, 'popular' revolutions may well target all
such dominant groups, not only for socio-economic, but also for ideological
reasons. Since the elites of various social groups often share similar forms of
education, media, clubs, friends, employment, and so on, and multiply
interact, even competitively, it may be assumed that such a dominant or
maybe better'elité ideology can also be shared by communication and
discourse.
Imposition and inculcation
The second question implied by this issue is whether the (shared) ideology —
or ideology fragments — of the dominant groups or elites can be somehow
Imposed' on dominated groups. This formulation of the question suggests
that dominated groups interiorize the 'dominant' ideology and accept it,
wholly or partly, as their own, whether or not that ideology is in their best
interests. Given the definition of power, domination and control in Chapter
16, this means that the elites are able (partly) to control the minds of the
dominated group. Since ideologies usually and largely are acquired through
discourse, and because the contemporary elites obviously control the means
of ideological reproduction, and especially the mass media, the question
essentially boils down to two interrelated empirical questions, namely,
whether the mass media mainly represent the ideologies of the elites, and
whether these ideologies have the intended influence on the ideologies of the
('dominated') public at large.
The first of these questions hardly needs to be further investigated: All
research shows that the ideologies that are most prominent in the media ar 2
largely those of the elites, and not of any dominated or oppositional groups.
Where moderate forros of oppositional ideologies (such as feminist or
environmentalist ones) have access to mainstream media at all, they are
consistent with those of significant 'fractions' of the dominant elites.
Dominant ideologies?
181
Also, this does not mean that the symbolic media elites (defined as senior
editors, leading reporters and columnists) always fully agree with for
instance political, business or academic attitudes and ideologies, let alone
with ah specific issues. As suggested, there are different interests and
attitudes. However, on fundamental issues, there is a rather broad consensus.
Thus, no mainstream Western media, nor other power elites are (today) anticapitalist, socialist, feminist, pacifist or anti-racist. Moreover, and even more
crucially, dominant elite groups have prominent access to the mass media.
Whether or not they are occasionally criticized (as corrupt politicians or as
polluting industries may), their overall representation is generally favourable
or at least respectful. I other words, through the media, other elite groups
and their discourses and opinions are at least able to reach the public at
large: They have an effective public voice.
The second question, about the ideological influence of the mass media, is
as complex as it is crucial in this debate. Much research suggests that the
general, ideological influence of the media is pervasive, especially in those
domains where media users have no alternative ideological sources or
personal experiences that are blatantly inconsistent with the dominant
ideologies as conveyed and reproduced by the mass media, as is typically the
case for ethnic ideologies or foreign policy ideologies? On the other hand,
much contemporary research emphasizes that even where such ideological
control takes place, media users are active and flexible and able to reject
persuasive ideological statements where necessary, or adapt such ideologies
to their own needs, interests or circumstances. Indeed, there are many
specific examples where pervasive ideolojical influence of the elites through
the mass media did not take place at a11.
In order to explore the implications of such apparently contradictory
empirical results, we need to know more about which dominant and
dominated groups are involved, what ideologies these respectively may
have, and under what conditions which dominant ideologies may be inculcated in which dominated groups. Again, these questions are not merely
conceptual but empirical, and their foil answers must hence be given in
detailed research into the ideologies of various social forrnations.
What are 'dominated' groups?
One flrst issue we need to deal with here is whether the very general and
hence fuzzy notion of 'dominated' groups, that is, as those dominated by the
elites, is a realistic construct. As long as we talk about socio-economically
defined'classes', as is the case in most traditional research and especially in
the Marxist tradition, the question may be slightly less complex, but
obviously ideologies in contemporary societies are not limited to classes.
The question then is whether these 'dominated groups' collectively have
or share (fragments of the same) ideologies. For the same social and
economic reasons as it was assumed that the elites must share ideological
182
Society
fragments, the non-elites must share such fragments, if only because of their
similar non-dominant position, and hence at least some shared interest,
namely, lack of power. Of course women, minorities, the poor, the workers
and so on, each have their group ideologies that provide the basic framework
that can account for their specific experiences in everyday life, for their
(dominated) position in society, and for possible forms of opposition, dissent
or resistance, that is, belief systems which Mannheim called'utopias',
because they formulate alternatives to currently dominating ideologies. 5 But
although there will be conflicts of interest (e.g. between white workers and
black workers, between poor men and poor wornen), their similar relations
to the elites suggest common ideology fragments which may lead to political
coalition formation, for example defining movements such as the Rainbow
Coalition in the USA.
Theoretically, there is no reason why these various non-dominant groups
would adopt dominant ideologies if these are inconsistent with their daily
experiences, their opinions about social events and their basic interests.
Indeed, if they were to do so, such ideologies would monitor group
knowledge and attitudes that would continuously clash with the daily
experiences of most group members. Thus, when confronted with (implied)
elite ideologies in the media, the public(s) at large would, again theoretically, only adopt those ideology fragments that are also in their own
interests and reject or ignore those that do not'fit'.
For the majority of the white public one prominent example may be the
adoption of fragments of racist ideologies, since these are also in their own
best interests. On the other hand, liberal market ideology fragments that
accept unemployment as a necessary aspect of capitalist production, or
promote a further destruction of the welfare state, may well be much less
broadly accepted, especially in the working class and the lower middle
classes. Following this argumentation, the general dominant ideology thesis
would in many instances be invalid, and would only apply for specific
ideology fragments and for specific non-dominant target groups, for example
whites, or men, or the middle class.
Whenever and wherever it does take place, ideological dominance may
take many forms and occur in different situations. Preventing solidarity
among non-dominant groups is a well-known and powerful device, namely,
of conquering the enemies by dividing them. Another strategy is to prevent
or mitigate group identification in the first place. We have already seen that
group identity and identification is a crucial implication of the acquisition of
ideological schemas throughout a group.
For instance, liberal socio-economic ideologies and especially their pervasive and persuasive genre expressions (in news, background stories,
advertising) in the mass media, may especially address media users as
individuals. In situations of social and economic crisis, thus, ingroup
solidarity among non-dominant classes may be prevented by suggesting that
each person 'can make it', as was the case in the conservative rhetoric of
Dominant ideologies?
183
'popular' capitalism of Thatcherism and Reagonomics, and the increasing
power of the New Right. 6
At the same time union membership may be discredited as 'communise,
radical' or simply out of date. Thus, forms of intra-group solidarity of
dominated groups may be prevented or obstructed. The same divisions may
be created among women by discrediting feminism; they may be discredited
among ethnic minorities by emphasizing ethnic crime or by discrediting
multiculturalism through allegations of political correctness, on the one
hand, and at the same stressing the positive role of the governrnent and the
'offering' of integrated minority lelp' by mainstream institutions, on the
other hand. Obviously, such strategies are not always successful, and
resistance and opposition may be able to challenge them in many ways, thus
leading to specific social changes, including some in the ideologies of
dominant groups. 7
Further complications
Of course, deeper analysis of there ideological processes is in order, because
the picture is much more complicated. For one, even within dominant
groups, there are ideological dissidents. That is, there are elite group
members (leading politicians, journalists, scholars, etc.) who reject and resist
dominant ideologies and may even 'side with' dominated groups, as has
been the case in most ideological revolutions. The converse is also true —
members of dominated groups may espouse elite ideologies, if only in order
to get, individually, recognition or access to other resources that the elites
will provide to them as tokens of their gratitude for their 'defectioñ .
Examples may for instance be found among some minority group members
who Nave espoused ideologies (e.g. about'political correctness') that are
clearly inconsistent with those of their own group (see Chapter 28).
Another well-known complication is the fact that, despite what has been
said aboye, there are cases in which elite ideologies are successful among
specific dominated groups even when they are inconsistent with the interests
of most group members, as is the case for neo-liberal market ideologies.
One, equally well-known, explanation for such success, apart from their
pervasiveness in the mass media and public discourse, and the social
processes of individualization and competition among the dominated groups,
are the various mechanisms of manipulation.
That is, what the public discourses of such ideologies typically do is to
tone down the obviously inconsistent parts of the ideology and emphasize
those parts that may be more attractive. Thus, racist (and some conservative)
parties may foment ethnic prejudices, blame immigrants or minorities for
social problems, and may thus attract lower (middle) class voters and
supporters. They will, however, seldom advertise their conservative policies
when it comes to the position of women and the consequences, for the poor,
of their market ideologies. At the same time, those who share more subtle
184
Society
versions of racist ideologies may publicly mitigate their racist attitudes, and
try to influence those among the population who reject overt racism, but who
may be sensitive to, for example, ecological or social ideas. In that case,
reference to overpopulation, scarce natural resources, or the cultural 'backwardness' (e.g. in the treatment of women) of some immigrant groups, may
be used as 'rational' arguments in favour of an immigration control that may
be acceptable even to liberals.
In the same way, the mass media will generally select or focus on those
'facts' that are consistent with elite interests, and vice versa, thereby
persuasively influencing the models, and indirectly the knowledge and other
social representations of the public at large, as described before. Most,
prominent are examples of nationalist war propaganda, and the public praise
of the blessings of the 'freedom' or the'flexibility' of the market, in which
the multiple negative consequences for large groups of the population will
be selectively obscured or simply ignored.
Strategies of ideological control
Thus, elite ideologies may well be adopted more broadly among the
population at large or among specific dominated groups under the following
conditions.
1 They are able to divide the non-dominant groups, by being at least
attractive to, or in the interest of, some non-dominant groups, thus
preventing intra-group solidarity and the organization of counter-power,
for example sexism and especially racism, thus preventing non-elite
solidarity and sharing of dissident ideologies, for instance among women
and minorities.
2 Preventing ingroup solidarity of (important) non-dominant groups by
creating divisions within the group and by addressing group members as
individuals, for example dividing women between 'feminists' and the
'others', or enticing lower-class members with liberal rhetoric of personal responsibility and upward mobility.
3 There are no (strong) popular alternatives to elite ideologies, or these
altematives, are unknown or marginalized, for example racism, because
anti-racism is virtually excluded from the mass media; or neo-liberalism
after the dernise of socialism and communism.
4 The elites (and especially media editors) prevent or limit the access to
public discourse of leaders of non-dominant groups (no feminist, antiracist or political 'radicals' in mainstream media), or will marginalize or
discredit them among the population at large or even among their own
groups.
5 Popular ideologies are seemingly adopted by the elites, but in a very
moderate way, thereby avoiding major conflicts with the interests of the
elites, for example environmentalism and — partly — feminism.
Dominant ideologies?
185
6 ff elite ideologies are largely inconsistent with relatively strong and
known ideologies of dominated groups, the elites have the special means
of media access and control, and discursive strategies of manipulation of
knowledge and opinions, for example by emphasizing the ideological
implications that are less inconsistent with the interests of dominated
groups, or de-emphasizing those that are inconsistent with these interests, for example nationalism, militarism, and especially neo-liberalism
and neo-conservatism.
Of course, there are other means of ideological control, but these cover a
wide variety of forms of ideological dominance. The more specific but
crucial discursive strategies involved in these forms of ideological domination will be discussed in more detall later.
Concluding remark
This discussion suggests that the argumenta for the dominant ideology
hypothesis are not very persuasive, but that in many situations and under
specific conditions it does seem to hold true. It is the task of a more detailed
theory of ideology to specify how and where it does apply, and where it does
not apply. It is, however, a very general and abstract thesis, and it is clearly
necessary that it be translated into the detailed structures of social cognition,
discourse, communication and social structures, before it can be evaluated
more rigorously. Despite the large ideological variety and confusion of
contemporary society, the evidence strongly suggests that, given the increasing control of the elites of the mass media, and the increasing role of the
mass media as the major means of ideological control of society, elite
ideologies will generally tend to be dominant, as defined aboye. Popular
ideologies may become dominant only (a) if they have broad support within
or across several dominated groups, (b) if leaders of such groups have access
to public discourse, and especially the mass media (which implies that at
least some mass media need to collude with them), and more generally (c) if
these ideologies are not fundamentally inconsistent with the interests of the
majority of the elites.
Organizing the reproduction of ideologies
In the analytical sequence that carnes us from the psychology of individual
cognition and action, and the microsociology of everyday situated interaction (including discourse), to the macrosociology of group relations,
power and shared belief systems, we fmally need to examine the role of
institutions that organize, manage or propagate such cognitions, actions,
interactions and group relations. In the discussion about the role of discourse
in the reproduction of ideologies, we shall further investigate how ideologies
are reproduced in and by the text and talk of families, peer groups, schools,
media, churches, unions, clubs, social movements, agencies, corporate
businesses, and so on. In the previous chapters, we have found that the
media play a central role in the reproduction of dominant elite ideologies.
Therefore, before I discuss the discursive details of such reproduction
processes, a sociological analysis needs to focus more generally on the
ideological role of organizations and institutions.
In many ways, institutions or organizations are the 'practicaP or social
counterpart of ideologies. That is, in the same way as ideologies organize
group cognition, institutions and organizations organize social practices and
social actors. Merely being a 'group' of women, journalists, teachers, or
anti-racists, may not be enough to organize members' actions effectively,
and to achieve desired group goals, either individually or jointly. Institutions
and organizations may co-ordinate common goals and actions, provide or
distribute resources and other conditions and constraints, elect or impose
leaders, and so on.
Similarly, in order to organize ideological practices, we may assume that
ideological institutions are needed. 1 In other words, ideological institutions are
created that (also) have as their task the 'realization' of a shared ideology.
Probably there are few institutions that are exclusively ideological, that is,
geared towards the propagation of belief systems only. Churches may still be
the most obvious example, although in practice, and in order to realize their
ideological goals, they also have several (other) social aims and activities,
such as welfare and community services. At a more basic levet, also families
and their socialization practices are pardy ideological, because of their role in
the socialization of norms, values and fragmenta of ideology. 2
Schools, universities and the whole education system are among the most
complex, elaborate and pervasive ideol6gical institutions, if only because
Institutions
187
they involve virtually afl members of society, intensively and daily, sometimes for more than twenty years. Geared mainly towards the reproduction
of knowledge and the acquisition of skills, they obviously also operate as
major means for the reproduction of the dominant ideologies of society,
although in sorne cases they also facilitate the propagation of counterideologies. Indeed, schools and especially universities are among the few
institutions where enough freedom (from state intervention, from the market,
etc.) exists for 'dissidents' to voice their opposed ideologies?
Despite this pervasive role of education, in contemporary information
societies much of the ideological work of the family, the Church and the
school is taken over by the mass media as an institution. While mainly
geared towards the production of information and entertainment, they at the
same time constitute the most complex institution for the public expression
and challenge of ideologies: Without the media, and given the reduced role
of the church, and the limitation of schooling to children and adolescente,
public debate about issues, and shared knowledge about what happens in
society and the world, would at present be unthinkable. It may therefore be
assumed that in the reproduction of ideologies, the media play a central role.
Social representations are easily and widely shared because of these forms of
accessible public discourse, and the same is true for the ideologies that
underlie these representations.
The structures, strategies and practices of these social institutions need not
only be oriented by practical reasons of organization, efficiency, the distribution of roles or resources or the attainment of goals. They may also reflect
and facilitate ideological concerns. Lessons, textbooks, exams, assignments,
corrections and sanctions in educational institutions, thus, may be organized
partly by ideologically based aims to teach and inculcate 'the right things',
including the'right' ideologies in the first place. In a less organized way, the
same is true for the various socialization discourses in the family.
The media
Though less explicit, but therefore probably more pervasive and influential,
the same is true for the media. The production of news, advertising,
documentaries, movies, games, talk shows and other shows, among many
— other.-media genres, may thus be examined in detail for the ways they
organize actions, discourses, sounds and images in such a way that ideological production and reproduction, including processes among the audience,
are most effective. In news gathering, such ideological concems monitor
assignments, beats, interviews, press conferences, press releases, selection
and decision procedures, among other practices. That is, these practices are
govemed by professional expertise and attitudes and ideologies about what
is true or false (fact or opinion), interesting or uninteresting, newsworthy or
not, relevant or irrelevant, and so on. News values are among the many
ideological systems that guide such practices — these specify, for instance,
188
Society
the preference for news about elites, negative events (especially those caused
by others), our own cultural group and world region, and so on.
13ut, less directly, the same is true for seemingly less ideological decisions
about who has access to the media, who is interviewed, covered, and who
will be quoted. It is well known that elite persons, organizations and states
are predominant in these pattems of access, and hence also the opinions and
ideologies of such elites. And sine most journalists in the West are white,
male, middle class and heterosexual (among other identities), it is most
likely that they will favour the access and the opinions of 'similar' news
actors. Most research confirms this assumption.
In sum, the routines, actors, events and institutional arrangements in
newsmaking are biased towards the reproduction of a limited set of
dominant, elite ideologies, as we have seen in a previous chapter. This is not
only true for news production, but also for current affairs programmes,
documentaries, shows, and other categories of media discourse.
What has been said about institutional production routines and constraints
is reflected in their products. Thus, preferential access is reflected in
preferential quoting, favourable opinions and hence style, access to the
opinion page, preferred topics, and in general in all aspects of media
discourse. For such complex ideological reasons, thus, we get more news
and opinions about alleged or socio-economically less destructive 'minority'
crimes than about real crimes of discrimination by employers or other elités,
more news by and about men and about topics that interest men more than
women, and so on. These are familiar research fmdings, and my point is
merely to recall them in order to illustrate the ideological conditions,
practices and products of institutions.
Crucially, the same is true for the consequences of such ideological
institutions for the reproduction of ideologies among the population at large,
as I have already discussed when examining the dominant ideology thesis.
Despite the personal differences and freedom of media users in their
processing and use of media discourse, the overall ideological effects of the
media are undeniable — the range of acceptable social ideologies is more or
less identical with those that have preferential access to the mass media.
Fundamental norms and values, the selection of issues and topics of interest
and attention (agenda setting), selective if not biased knowledge about the
world, and many other elements or conditions of ideological control, are
presently largely due to the mass media, or indirectly to the groups and
institutions, such as those of politics, that have preferential access to the
media. Of course there will be debate, opposition, differences of opinion, as
well as differences among newspapers. However, these are well within the
boundaries of tolerable ideological variation. No serious newspaper advocates, for instance, the abolition of the market, the abolition of all arms and
armies, a total reversal of all gender roles, so that women will be put in
charge of the world and all major institutions, let alone the control of the
mass media by independent monitoring organizations that will evaluate their
truthfulness, quality, and the total absence of gender, class, ethnic or other
Institutions
189
biases. In sum, within a theory of ideology, the pervasive role of ideological
institutions such as those of politics, education and especially the mass
media explains the very social conditions of ideologies, namely, the means
and the ways of their being shared by large numbers of people and groups in
the first place.
Institutional racism
Taking up again the example of racism, we should ask how racist ideologies
are sustained and reproduced by institutions and organizations. The most
obvious example in most countries in Europe and other white-dominated
countries is the presence and activities of racist political parties. 5 Although
politically nowhere dominant beyond the local level of some neighbourhoods and cides, and although often marginalized by the mainstream media,
their indirect ideological influence is considerable. Even when covered in a
context of conflict, for instance by quoting provocative statements of their
leaders, or by highlighting counter-demonstrations and protests, they are as
widely known as their ideologies are. The radical versions of these ideologies may be generally rejected by the elites, but it has been often observed
that more moderate versions of their xenophobic or anti-immigrant slogans
have found wide currency and even support among mainstream parties, as
has been the case for instance for the conservative parties in the USA, the
UK, France, the Netherlands, Germany, Austria and Italy, among other
countries.
Increasingly harsh immigration restrictions, earlier advocated only by
racist parties, are now standard government policy everywhere. The same is
true for various policies that turn back (or never introduce) the gains and
claims of the civil rights movement or similar movements in other countries.
Popular support for such policies among large sections of the white
population is guaranteed, after the ideological onslaught of racist and
conservative propaganda, which tends to blame many social ills on the
presence or the activities of immigrants and minorities. Immigration can
thus easily be targeted as one of the major causes of unemployment,
diminishing welfare or the real or alleged increase of crime. The mass
media, and especially the conservative popular press, play a crucial role in
the persuasive support and propagation of these ideologies.
And although I focus here on the production and reproduction of
ideologies, it hardly needs to be added that such ideologies also sustain
concomitant social and political action. Ideologies are translated finto actual
policies, which are executed in concrete practices, for example of the
immigration services, the police or the courts, or the media. The negative
examples of the elites and state agencies are followed, often more blatantly
or even violently, by organizations or youth groups who openly discriminate
or attack immigrants and minorities. In a few dornains in society, the
institutional and elite propagation of ethnocentrist, xenophobic and racist
190
Society
ideologies clearly and directly influences everyday practices of exclusion,
marginalization, problematization and violence directed against others, as is
the case in the anea of ethnic relations. Colonialism, slavery, segregation,
Jim Crow, the Holocaust, and at present Rwanda, Bosnia and South Asia,
are the well-known illustrations of that observation.
In sum, racist ideologies, and especially their popular and populist
versions, are sustained by a large number of important institutions and
organizations. Extremist right-wing parties, conservative parties and thinktanks, the popular press, phone-in radio, racist pamphlets, marginal but
influential racist scholars and their publications, are among the many
institutional factors that are involved in this reproduction process.
Again, although the radical versions of these ideologies may not be
predominant, moderate versions may well have become dominant in Western societies where conservative forces are in the majority. Even leftist and
social democratic parties and organizations doñ t escape the pressures of the
broad popular (white) support for such ideologies, and adapt their ideologies
and policies accordingly. This does not only show in the support of antiimmigration or anti-minority attitudes, but especially also in the marginalization of anti-racist groups and ideologies.
Indeed, one of the main problems in Western societies may not be that
moderate racist ideologies are influential, but rather that the official nonracist norm, as established by law and constitution, is not institutionalized in
such a way that such ideologies are energetically combated. There are antiracist groups and institutions, but these are minor and often have a bad press
or little support among the population at large, as well as among the elites.
They may be officially marginalized as much as the extremist-right, while
allegedly being too'radical'. In such a political evaluation, thus, both racism
and anti-racism are rejected, thus leaving a vast consensus intact in which
anti-immigrant ideologies may flourish because they are simply deemed not
to be racist, but commonsense. We shall later see how political and media
discourse constructs and sustains such a broadly organized consensus of
white domination.
Part III
DISCOURSE
w
The Relevance of Discourse
The special relevance of discourse
In the tbird part of this study I fmally focus on another crucial dimension of
ideology, namely, its expression and (re)production in social interaction in
general and in discourse in particular. Having assumed that ideologies are
shared social representations that have specific social functions for groups,
we need to fmd out how such ideologies are acquired, constructed, used and
changed by social group members. This means that, after the excursion into
the social macro dornain of groups, group relations and institutions, we now
need to get down to the micro level again, that is, to the level where
ideological production and reproduction is actually being achieved by social
actors in social situations.
Against the background of the classical approach to ideology, such a
micro-level study of interaction and discourse is especially relevant. Not
only does the traditional account of ideologies tell us litde about the precise
nature of ideologies (namely, as mental representations), but it is also
unspecific about how exactly ideologies come about, and what role social
actors play in their construction and reproduction. This also means that such
approaches largely ignore how a typical macro notion such as ideology
should be related to typical micro notions such as actors, actions, social
practices, discourses and social situations.'
By focusing especially on the role of discourse in the reproduction
processes of ideologies, I do not imply, as some current approaches do, that
1 reduce ideologies, or their study, to discourse and discourse analysis. 1
Discourse, language use and communication do play a special role in such
processes of reproduction, but ideologies are also being expressed and
reproduced by social and semiotic practices other than those of text and talk.
From the study of racist and sexist ideologies, for instance, we know that
many forms of non-verbal discrimination aso exhibit ideological beliefs.
Besides these well-known practices of discrimination, also other semiotic
messages (e.g. photographs and movies) may of course express underlying
ideologies? When social members observe and comprehend such (non-
192
Discourse
verbal) practices, they may also infer underlying opinions of actors; and
across contexts these may also be generalized to more abstract underlying
social attitudes and ideologies. They may do so through an inferential step
that tells group members: 'This apparently is how we do it' or, 'This is
apparently the way to deal with members of such and such a group.' In sum,
although discourse is often crucial in the expression and reproduction of
ideologies, it neither is a necessary nor a sufficient 'medium' of reproduction.
Although this part of the study focuses on discourse, we should bear in
mind that it is paradigmatic for a broader study of ideological practices in all
domains of society, from non-verbal communication to the myriad of other
social actions and interactions that define everyday life. Also, we should not
forget that discourse is often embedded in or otherwise related to such nonverbal interactions, as is the case for talk and text at honre, in parliament, in
school, in the newsroom, the workfloor, the office, the shop, the agency, the
hospital, the police station or in prison. Ideologically based dominance and
inequality, conflict and competition, resistance and opposition, as discussed
before, thus, are implemented and reproduced in many ways, both discursively and in other interactions.
Discourse, however, has a special status in the reproduction of ideologies.
Unlike most other social practices, and in a more explicit way than most
other semiotic codes (such as photos, pictures, images, signs, paintings,
movies, gestures, dance and so on), various properties of text and talk allow
social members to actually express or formulate abstract ideological beliefs,
or any other opinion related to such ideologies. Specific actions only allow
more or less indeterminate inferences about the underlying opinions of
actors, but as such cannot express general, abstract or socially shared
opinions.
With visual messages this is somewhat easier, and in some cases more
effective than through discourse. But in general, there is no semiotic code as
explicit and as articulate in the direct expression of meanings, knowledge,
opinions and various social beliefs as natural language (and of course in
various sign languages). If an image is worth a thousand words, this is
mostly because of the visual details that are hand to describe verbally. This
means that images may be particularly apt at expressing the visual dimension of mental models. If images express opinions or general beliefs and
ideologies, they do so rather indirectly, and hence are in need of (indeterminate) interpretations. This does not imply that, in communication, such
indirect expressions of opinions and ideologies need to be less persuasive.
On the contrary, a dramatic photograph of a specific scene, event or person
may be a much more 'powerfur means of expressing opinion than words.
However, this persuasiveness is precisely based on the concreteness of the
'example', and needs reader-based inferences about what the picture actually
'means', as is also the case for model-based storytelling or other examples
used as a means to convey abstract opinions and ideologies.
The relevance of discourse
193
Discourse enables social actors to formulate general conclusions based on
several experiences and observations. It is able to describe past and future
events, it is able to describe and prescribe, and may describe actions and
beliefs at any level of specificity and generality. And for us, most interestingly, discourse not only exhibits ideologies indirectly, as other social
practices may do too, but also explicitly formulates ideological beliefs
directly.
Thus, in many situations of intra- and inter-group text and talk, social
members are able to tell or remind others or novices about the ideological
beliefs shared by the group. Ideological socialization, therefore, largely takes
place through discourse. In interactional confrontations with members of
other groups, people are similarly able to discursively explain, defend or
legitimate their ideologies. In other words, discourse allows direct and
explicit expression of ideologies, but the crucial function of such (usually
generic, general) expressions is in their social consequences, namely, the
acquisition, change or confirmation of social beliefs. 4
In this and the next chapters, I shall describe some of the dimensions of
the relations between discourse and ideology. This investigation is merely
illustrative — many volumes can be written about the many ways ideologies
are expressed in text and talk My approach here is primarily conceptual and
theoretical. I want to know, more generally, how discourse expresses or reproduces underlying ideologies, and not study specific ideologies or specific
language or discourse structures (such as topics, pronouns or metaphors). In
a later study I hope to focus in more detall on the role of discourse structures
in the reproduction of ideologies.
The concept of discourse
In order to understand how ideology relates to discourse, let me first
summarize my discourse theoretical framework, especially sine this is
somewhat different from others that study both discourse and ideology, such
as the more philosophical approach by Foucault. 5 As indicated before, my
approach is essentially multidisciplinary, and combines an analysis of
linguistic, cognitive, social and cultural aspects of text and talk in context,
and does so from a critical, socio-political perspective. b
The concept of discourse used here is just as general, and hence as fuzzy,
as that of language, cotnmunication, society or, indeed, that of ideology.
Although its 'definitioñ is the task of the whole discipline of discourse
studies, a few remarks are in order about my use of the term'cliscoursé in
Chis study. This is also necessary because in many current studies of
ideology and its relations to discourse, other (sometimes confusing) discourse concepts are used. 7
Communicative events versus verbal products
The primary meaning of the terco 'discourse' as it is used here, and as it is
now generally used in more socially oriented discourse analysis, is that of a
194
Discourse
specific communicative event. Such a communicative event is itself rather
complex, and at least involves a number of social actors, typically in
speaker/writer and hearer/reader roles (but also in other roles, such as
observer or overhearer), taking part in a communicative act, in a specific
setting (time, place, circumstances) and based on other context features. This
communicative act may be written or spoken, and, especially in spoken
interaction, usually combines verbal and non-verbal dimensions (gestures,
face-work, etc.). Typical examples are an everyday conversation with
friends during dinner, a dialogue between doctor and patient, or writing/
reading a news report in the newspaper. We may call this the extended
primary meaning of the term 'discourse'.
In the everyday practice of discourse studies, however, we often also use
a more restricted primary meaning of 'discourse.' In that case, we abstract
the verbal dimension of the spoken or written communicative act of a
communicative event, and usually refer to this abstraction as talk or text.
That is, in this sense 'discoursé is rather being used to refer to the
accomplished or ongoing 'product' of the communicative act, namely, its
written or auditory result as it is made socially available for recipients to
interpret. 'Discourse' in that case is the general term that refers to a spoken
or a written verbal product of the communicative act.
In earlier text linguistics, and still among some discourse linguists, a
related distinction is made, between 'discourse' and 'text'. Here 'discourse'
is used to refer to the actual, socially displayed text or talk, and 'text' to its
abstract (e.g. grammatical) structures. This distinction implements for discourse analysis the well-known distinctions between langue and parole or
between competence and performance in structural and generative linguistics. 'Discourse' is then a unit of language use or performance (parole), and
'text' an abstract theoretical unit (like a noun phrase, clause or sentence) that
belongs to the realm of abstract linguistic knowledge or competence or to
the system of the language (langue). Although relevant, I no longer use this
distinction. In contemporary, multidisciphinary discourse analysis it has
become either too confusing or obsolete — discourse studies now generally
analyses discourses as forms of language use. Such a focus on concrete,
ongoing language use does not mean that the theoretical account itself is less
abstract. In the same way as linguists abstract grammatical properties from
actual verbal utterances, discourse analysts do so when they describe, for
example, gestures, intonation, pauses, repairs, graphical design, narrative
structures, metaphors, turns, closing sequences, and so on.
Tokens versus types
Whether in its extended or restricted meaning, namely, as talk/text or as
complex communicative event, 'discourse' in this primary meaning is used
to refer to particular objects or tokens, that is, to unique occurrences
involving particular social actors in a particular setting and context. This
uniqueness is for instance defined in terms of the unique combination of
The relevance of discourse
195
these words, intonation, gestures, meanings or acts being accomplished now
by these participants. To mark this specific use of the notion of 'discourse',
we use indefinite or definite anides or demonstratives: we speak about 'a
discourse', 'the discourse' or 'that discourse'. That is, 'discourse' is a count
noun here.
In the age of printing, xeroxing and computer files, copies may be made of
the spoken or written expression of such a unique discourse, for instance on
tape or in a book or newspaper. But even then we say that these are copies
of (the expression of) the 'same' discourse.
As elsewhere, there are the usual delimitation problems: Where does one
discourse end and the next one begin, for instance in a sequence of
conversations, or in a collection of printed texts, for instance in a newspaper,
book or encyclopedia? Are the different instalments of an anide, a TV film,
or a daily story, one or more discourses, even when they are physically noncontiguous in time or place? There are many examples where there is an
ambiguity between such discontinuous instalments of the 'same' text or talk,
on the one hand, and sets of 'intertextually' related discourses, on the other
hand. Indeed, whereas a continuous spoken dialogue is usually considered as
representing one discourse, a written dialogue or debate is rather seen as an
intertextually related sequence of texts, even when it may be called 'one'
debate in both cases.
This is not the place, however, to solve such well-known problems of
delimitation and definition. To simplify matters, I simply follow commonsense practices here, and speak about one dialogue when it is continuous
in time (not in space, because participants may talk to each other over the
phone), has the same participants, and has a marked beginning and end. And
for written texts we assume that they have the same writer(s), have marked
beginning and end, and usually, though not always, that they are physically
continuous (exceptions are, indeed, several instalments of the 'same text'
appearing at different times, or separate parts of the same time appearing in
different locations of the same medium (e.g. a front page story in the press,
continuing on an inside page). Both for spoken and for written discourse, we
usually further require that they are globally coherent, that is, that they form
a meaning unit, and not only a physical unit of continuous expression. But
this requirement is itself problematical for everyday conversations that are
characterized by several unrelated topics, or for instance literary texts, like
poems, that do not seem to have an obviously unitary, global meaning.
These problems and examples also show that 'discourse' is a highly
complex and ambiguous notion, and that as soon as we really want to give a
'clefinitioñ we already need to start making all kinds of analytical distinctions, use other concepts, and indeed start to theorize about discourse. Hence
it is usually a rather pointless exercise to give exact definitions. As
suggested aboye, thus, discourse is as general and therefore as vague a
notion as 'languagé , 'society' or 'culturé .
Besides the specific (extended or restricted) notion of'cliscoursé, there is
also a more abstract concept. Instead of specific, unique tokens, we may also
196
Discourse
use 'discourse' to refer to abstract types. Thus, instead of referring to this
particular conversation, story or news report, we may also use the notion of
discourse in order to designate conversations, stories or news reports in
general. When we make theoretical, that is, general, assertions about
discourse, they are of course about types, not about tokens. We may say that
'a' or 'the' news report or story consists of a number of conventional
categories, such as an initial summary (e.g. a headline and a lead) or a
concluding coda. That is, in this case we characterize a potentially infinite
set of real or possible tokens that satisfy such properties. This abstract notion
of discourse may similarly be restricted as well as extended. We may refer to
a dialogue as the verbal result of a communicative event, or to the whole
communicative event. In this chapter, we only talk about discourse and its
properties in general, not about particular instances or tokens of text or talk
as we would do when analysing concrete examples.
Text and talk of social domains
To make things even more complicated, there are at least two other main
meanings of the concept of discourse. First, closely related to the notion of
discourse referring to an abstract type, the concept may be used to refer to
specific genres, mostly in combination with an adjective denoting a genre or
social domain, as in political discourse, medical discourse and academic
discourse. In this case, the notion of discourse is also general and abstract,
but selects a specific set of (abstract) discourses or genres. Thus, political
discourse may be the overall designation for all discourse genres that are
used in the realm of politics, or the discourses used by politicians, and so on.
In this sense, 'cliscoursé is not simply a specific genre (like a parliamentary
debate or a propaganda leaflet), but rather a socially constituted set of such
genres, associated with a social domain or field.
Finally, we may distinguish an even more abstract and higher-level notion
of discourse. Instead of referring to afi the text and talk, or the discourses of
a specific period, community or a whole culture, we may also use the very
abstract and generic notion of the 'discourse' of that period, community or
culture — including all possible discourse genres and all domains of
communication. Other notions sometimes used here are discourseformation
or discursive formation, and order of discourse, following sociological uses
of the terms'social formatioñ and 'social order', respectively. Depending
on oné s theory of discourse and society, also this highly abstract notion of
discourse may be restricted (all text and talk) or extended (all communicative events, including language users, contexts, etc.). It is this last, very
abstract and general notion of discourse that is often related to the equally
general, abstract, social and shared notion of ideology. Indeed, this notion of
discourse is sometimes even collapsed with that of ideology, a practice of
reduction that I rejected as theoretically, empirically and analytically misguided.
LI
The relevante of discourse
197
Confusion here is even worse when this broad, philosophical concept of
discourse also includes the ideas and ideologies of a specific period or social
domain. As is of course often the case, the most general and ill-defined
concepts may sometimes become most popular. After all, in cultural fads
and fashions, ambiguity, myth and vagueness are often more attractive than
conceptual precision. This is currently also the case for manr postmodem
uses of 'discourse' in the humanities and the social sciences.
Whatever the ambiguities and fuzziness of the various notions of discourse introduced aboye, most share verbal (and related other semiotic)
properties. That is, I do not use the word 'discourse' (or 'text' for that
matter) for social structurés, interactions or communicative events that do
not have (also) a verbal character. Thus, societies, (sub)cultures or social
practices will not be described as discourses or texts here, even when they
may need understanding or interpretation, or when they are routinely
'accomplished' much like discourses.
Other semiotic 'discourses'
Finally, another well-known case comprises 'messages' in other semiotic
codes, such as (sequences of) images, movies, a dance and so on, especially
when these also have a verbal dimension. 9 I shall, however, limit myself to
commonsense notions here, and again only use the restricted notion of
'cliscoursé (text or talk) when referring to the verbal dimension of communicative interaction. Obviously, the extended notion of discourse, when
referring to a whole communicative event, may well also feature other
(visual, gestural) dimensions of communication and interaction, sometimes
closely intertwined with the verbal aspect, as is the case in spoken movies
and advertising. The only problem is that there is no everyday word to refer
in general terms to either integrated (verbal/non-verbal) 'discourses', or to
exclusively non-verbal semiotic 'messages', except by their specific words,
such as 'picture', 'photo', 'movié or 'advertisement'.
I do not use the semiotic terms 'signs' (or indeed 'signifier' or 'signified')
here. For discourse analysis these have generally become obsolete after more
than thirty years of increasingly sophisticated linguistics and discourse
studies. These notions were useful in early semiotics in order to describe, in
the tercos of early structural linguistics, some properties of non-linguistic
semiotic codes or objects, such as stories, movies, non-verbal sign systems,
or other cultural artefacts. Moreover, the notion of 'sign', following early
structuralism, is mostly used to denote minimal meaning units (like words),
and not maximal meaning units like whole discourses or movies.
Where necessary, I shall simply speak of non-verbal discourses, or use
specific genre designations. As is the case for other, more sophisticated
disciplines (such as linguistic, logic or communication studies), it is hardly
relevant to keep using traditional semiotic terminology to describe discourse
structures. However, as long as the study of other semiotic practices does not
have its own theoretical terminology, the integrated description of verbal and
198
Discourse
non-verbal 'messages' may still make use of such semiotic terminology.
This is especially so if such semiotic descriptions go beyond the mere
identification of isolated signs, signifiers or signifieds, and focus on more
conWlex structures of expression (signifiers), meaning (signifieds) and
use.
The study of discourse
Discourse studies, as it is understood in this book, is a cross-disciplinary
fleld of research that has emerged, especially since the mid-1960s, in
virtually all disciplines of the humanities and the social sciences. Initially
developed in linguistics, literary studies and anthropology, it soon also
spread to sociology, psychology, communication research and other disciplines. In principie, discourse studies as a separare cross-discipline besides
linguistics (or semiotics for that matter), would not have been necessary if
linguistic theories had paid attention to the study of actually occurring text
and taik in the first place. However, most hard-core linguistics focused on
grammar and on isolated sentences, even if there are directions of research
that may focus on the textual or interactional 'functions' of grammatical
structures of sentences. Hence, together with such other cross-disciplines as
socio-linguistics, pragmatics and the ethnography of speaking, discourse
analysis focuses on the systematic account of the complex structures and
strategies of text and taik as they are actually accomplished (produced,
interpreted, used) in their social contexts.
As suggested aboye, such a brief characterization of what I understand by
'discourse studies' (or the less adequate, but better-known term 'discourse
analysis') is relevant in order to distinguish Chis field from (some) more
impressionistic studies of discourse, especially in philosophy and literary
studies. Discourse studies of course focuses on the broad social and cultural
functions, conditions and consequences of text and taik, including, in our
case here, the role of discourse in the study of ideology. However, more
specifically, discourse and conversation analysis will typically always also
focus on systematic, detailed and theory-based analyses of actually occurring structures of text and taik. Thus, a mere paraphrase or summarization of
the 'content' of discourse, as also language users often do on the basis of
their knowledge of discourse, is usually found not to be a form of discourse
analysis in the sense intended here.
In its thirty years of existence, discourse studies has developed into a quite
sophisticated discipline, and it would be no serious contribution to our
insight into discourse (or ideology) if we were to simply ignore the many
advances in the many ateas of this new discipline.
However, given the ambiguity of the term 'discourse', we may expect the
same for 'discourse analysis', and there are therefore many directions and
approaches of research, and many fields of inquiry. Besides linguistic
(grammatical) studies of discourse, thus, we may find pragmatic studies of
The relevance of discourse
199
(speech) acts, conversation analysis, stylistics, rhetoric, or the sociolinguistic study of discourse variation in its social context. Most of these
studies focus on the various structures or strategies of text and talk, to be
discussed in the next chapter. However, also the psychology of discourse
production and understanding should be included in a broad, multidisciplinary discipline of discourse. The same is true for the study of microsocial
dimensions of interaction and contexts, in which relations between discourse
structures and, for example, properties of participants are being theorized.
In other words, the field of discourse studies as a discipline obviously
follows the study of text and talk in the various disciplines in the humanities
and the social sciences, and now also includes social psychology, communication research, political science and history. Ideally, an integrated study
integrates the analysis of discourse structures per se with the account of their
cognitive, social, political, historical and cultural functions and contexts. It is
in this broad, integrated and multidisciplinary approach that I locate the
study of the discursive expression and reproduction of ideologies.
21
Discourse Structures
On levels, structures and strategies
Typical for a discourse analytical approach to ideologies and their reproduction is that ideologies are not símply related to undifferentiated forms of text
or talk, but mapped on to different levels and dimensions of discourse, each
with its own structures or strategies. These various properties of discourse
are the result of theoretical analyses and therefore may vary widely in
different approaches.
Thus, conversation analysts exclusively focus on spontaneous, everyday
dialogues, linguists on the grammatical structures of discourse, whereas
pragmatics focuses on more speciflc properties of action and interaction,
such as speech acts, illocutionary force or politeness strategies. Whereas
earlier 'text linguistics' in practice tended to study mostly written texts, most
other contemporary approaches, especially in the social sciences, have a
preference for the analysis of spoken discourse, sometimes with the implicit
assumption that 'natural' language use is essentially oral and interactive.
Psychology on the other hand favours the study of (written) text comprehension, probably also because this is easier for experimentation in the
laboratory.
It needs little argument, however, that both spoken and written/printed
forros of discourse are the object of discourse studies, and that diere is no
more or less 'natural' priority here, at least not for all cultures that have
writing systems. Any approach that uniquely associates ideologies or social
representations with the interactive, face-to-face social construction of
'meanings' is therefore by definition incomplete: ideologies are also
expressed and reproduced by written text. Indeed, when it comes to the
mass-mediated reproduction of ideologies in contemporary society, face-toface interaction may even play a less prominent role than textual or onesided spoken/visual communication by newspapers and television.
From the sprawling cross-discipline of discourse studies that has emerged
from anthropology, sociology, linguistics, psychology and other disciplines
in the humanities and social sciences, we may hardly expect anything else
but a large variety of approaches, theories, methods and their underlying
philosophies. In order to give some background to the chapters that follow,
let us briefly summarize some of the main structures usually studied in
discourse analysis. At the same time, I give a brief indication of the ways
ideologies may impinge on such structures during their communicative
Discourse structures
201
manifestations. Note, though, that these indications are merely illustrations.
A proper discourse analysis of ideological expressions of course would
involve a much more detailed and systematic account of relevant structures
and strategies.
Graphics
Neglected in virtually all approaches of discourse studies, and obviously
irrelevant for the study of spoken dialogue, graphical structures of written or
printed text are literally a prominent, while actually visible, property of
discourse. Apart from some semiotic work on images or textual graphics,
theory-formation in this field is still scarce, and analyses hardly go beyond
impressionism. Yet, litde theory is necessary to understand that variations of
graphical prominence may constitute a crucial element in the expression of
ideologies. Whether a news report appears on the front page or on an inside
page of the newspaper, high on the page or at the bottom, left or right, or
whether it has a small or a banner headline, is long, short or broad, that is,
printed over several columns, with or without a photograph, tables, drawings, colour and so on, are all properties of the graphical representation of
just one genre that may have a serious impact on the readers' interpretation
of the relevance or newsworthiness of news events. Many advertisements are
inherently associated with images, colours and other graphical elements, and
sometimes lack verbal text altogether. The visual element of TV programmes is crucial, and also includes special discourse graphics. Modem
textbooks have a graphical layout that is assumed to raise and keep the
interest of children and adolescente. And so on for a large variety of other
written or printed genres.'
Graphical structures may have several cognitive, social and ideological
functions. Cognitively, they control attention and interest during comprehension, and indicate what information is important or interesting, or should be
focused on for other reasons, and may therefore be better understood and
memorized. They may signal communication forms and genres, such as the
difference between a news report and an editorial in the press, or between
theory and assignments in a textbook. Socially, graphical structures, including photographs, have a large domain of associations, for instance with
groups, organizations and subcultural styles, as the difference between a
popular tabloid and a serious mainstream broadsheet shows, or the type of
advertising in fancy magazines, street billboards, the subway or a supermarket leaflet.
At all these levels the possible expression of ideologies is obvious, for
instance through the graphical emphasis of positive values with ingroups,
and negative values with outgroups. Through images, photos, text placement, page layout, letter type, colour and other graphical properties, thus,
meanings and mental models may be manipulated, and indirectly the
ideological opinions implied by them. A serious theory spells out what
graphical structures exactly may have which of these various functions?
202
Discourse
Sound
The phonetic and phonological expression structures of discourse (the
'sounds'), though systematically studied since the beginning of modem
linguistics and phonetics, have also been neglected in discourse analysis. 3
Articulation or auditory reception or phonemes may be marginal to a typical
discourse analyst who prefers to look at discourse structures beyond those of
words, phrases or sentences. Yet, pitch, volume and intonation are a rich
source of variation by which, as in graphical expressions, emphasis, prominence or distinctiveness may be controlled as a function of semantic and
ideological importance and relevance, as well as of opinion, emotion and
social position (as in authoritarian commands versus polite requests). Since
most conversation analysts work with transcripts, precisely these 'sound
structures' tend to be partly ignored in analyses, or reduced to rather crude
forms of representation or description, with the exception of the study of
applause in public address.
Especially interesting for ideological analysis is the fact that subtle sound
variation may directly code for underlying opinions in event and context
models, that is, without explicit semantic articulation: Admiration, praise,
derogation, blame and many other functions of discourse may thus be
signalled implicitly — and hence deniably — as a function of ideological
beliefs. The sound structures of talk to or among women and men, whites
and blacks, superiors and subordinates, and generally ingroup and outgroup
members, may thus display, emphasize, conceal or persuasively convey
ideologically based opinions about events or the participants in the context.
Morphology
The study of word-formation is not exactly a main focus of concern in most
types of discourse studies, and usually associated with traditional sentence
grammatical research. Since stylistic variation, compared with other levels
of utterances, is limited here, the ideological impact on the way words are
formed in text and talk seems to be marginal, especially in languages that do
not allow compounds. Where relevant, for instance in the study of neologisms, such ideological effects usually will be studied in lexical stylistics.
Syntax
On the other hand, the study of sentence forms, syntax, has drawn attention
from (critical) linguists interested in ideological analysis from the start. 5
Variatioñ in the order or hierarchical relations of the structures of clauses
and sentences is a well-known expression of dimensions of meaning as well
as of other underlying semantic and pragmatic functions. Thus, order and
hierarchical position may signal importance and relevance of meanings, and
Discourse structures
203
may thus play a role in emphasizing or concealing preferred or dispreferred
meanings, respectively.
Agency and responsibility for actions may similarly be emphasized or deemphasized, for example by active or passive sentences, explicit or implicit
subjects, as well as word order. It needs little analysis to show that such an
important function of syntactic variation may have an impact on the
description of ingroup and outgroup actions, and hence on ideological
implications of text and talk. Position and role of clauses may signal
implications and presuppositions, which are closely related to what language
users should or should not know, and hence to 6 the ideological discursive
functions of exposing or concealing information.
Among many other features of syntax, pronouns are perhaps the best
known grammatical category of the expression and manipulation of social
relations, status and power, and hence of underlying ideologies. Ingroup
membership, outgroup distancing and derogation, intergroup polarization,
politeness, formality and intimacy, and many other social functions may thus
be signalled by pronominal variation. Ideologically based respect to others
may be given or withheld by using familiar or polite pronouns of address, as
in French tu and vous and Spanish tu (or vos in some Latin American
countries) and Usted. Given the group-based nature of ideologies, group
polarization and social struggle is thus prototypically expressed in the wellknown pronominal pair of Us and Them. Indeed, there are few words in the
language that may be as socially and ideologically loaded' as a simple we.
The close relationship between group identity, identification and ideology,
as discussed before, explains this particular function of this pronoun.
The specific set of choices that are made among the possible structures of
syntactic forro in one specific discourse, is usually called the (syntactic) style
of that discourse. Combined with lexical variations in the choice of words
(lexical style, see below) such syntactic style is often studied in a separate
domain of discourse analysis, namely, stylistics. Style may generally be
described as the overall result of the consistent use of variable grammatical
structures as a function of properties of the context (or rather of the
interpretation of the context as represented in context models). This means
that style is by definition a function of the ideological control of such
context models, as we have seen for the example of polite or impolite uses of
forms of address.
Semantics
Graphics, sound and sentence forms are usually categorized as 'observable'
expressions of discourse, traditionally called 'surface structures' in generative grammar. In some critical and ideological studies (often but not only
in a Marxist tradition), such structures may even be called 'material',
although, as suggested before, there is very little 'material' in abstract
204
Discourse
structures (one reason I used 'observable' between quotes). Yet, in a
somewhat sloppy but practical sense, we may say that surface structures are
the kinds of things that are actually and 'observably' expressed, shown and
displayed for interpretation by recipients. But we should remember that also
these 'observable' structures of expression are in fact abstract or mental
structures being assigned, by theorists as well as language users, to the
various physiological, auditory or physical (phonetic, printed) properties of
communication.
The meaning of 'meaning'
Undoubtedly crucial in all ideological analyses of discourse are the meanings expressed by or assigned to surface structures by discourse participants.
Unfortunately, there are few notions in the study of language and discourse
that are so complex and vague as that of meaning. Especially also in critical
or ideological studies, the notion is sometimes used so broadly that it has
lost virtually all 'meaning'. That discourse expresses, conveys, has, constructs and does many other things with meaning is both commonsense and
scholarly knowledge. Yet, we need a sophisticated semantics, or even
various types of semantics, to be able to spell out how exactly, what kinds of
meanings, are involved here. Simply talking about the'production of
meaning', as is usual in much contemporary critical studies, does not tell us
much about the role of discourse or ideology in communication, interaction
and society.
Thus, it cañ t hurt to recall old linguistic, philosophical and logical
distinctions between (conceptual) meaning or intension, on the one hand,
and reference, that is, as a relation between expressions and things being
referred to, denoted or talked about (i.e. the referents, denotata or extension),
on the other hand. Similarly, in an abstract analysis, it also makes sense to
distinguish between word or sentence meanings, utterance meanings, speaker's meanings, hearer's meanings and socio-cultural meanings (including
ideological meanings).
As is the case for all structures of discourse, all these different 'meanings'
result from different theoretical approaches. In traditional linguistics as well
as in common sense, words are associated with (word) meanings, as is still
the case in dictionaries. In structural and generative grammars, meanings of
sentences are formally constructed as a function of the meanings of words
and syntactic structures. In philosophical logic, meanings are abstract
functions that make sentences true or false, or that pick out referents or
extensions (objects, properties, facts) in some situation or possible world.
Meaning and interpretation
In the philosophy of language as well as in psychology and most of the
social sciences, meanings are not so much abstract properties of words or
expressions, but rather the kinds of things language users assign to such
Discourse structures
205
expressions in processes of interpretation or understanding. This also allows
for contextual variation: a speaker and a hearer may assign (intend, interpret,
infer) different meanings to the same expression, and indeed, the same
expression may therefore also mean different things in different contexts.
Hence, meanings of discourse or language in use are contextual or situated,
and depend on the (interpretation of the) participants.
Psychologists will then further spell out how such meaning assignments
or interpretations take place mentally, and what memory representations
(such as models or knowledge) are involved in meaning production and
understanding. Socially oriented discourse analysis will usually ignore such
cognitive'processing' of meaning, and focos exclusively on the interactive
or social construction of meanings in or by discourse. In this case meanings
are usually inferred intuitively by the analyst, and are not further analysed. It
is on this (rather shaky) basis that much ideological meaning analysis often
takes place.
As we shall see later, discourse meanings are the result of selecting
relevant portions of mental models about events. That is, knowledge about
events is thus mapped on verbaily expressed meanings of text and talk, and
hence partly constrained by the possible word and sentence meanings in a
given language or culture. Since models embody opinions, which may in
turn have an ideological basis, also the meanings that derive from such
'ideological' (biased, etc.) models may embody ideological aspects.
Many of these opinions may be conventionalized and codified in the
lexicon, as the respective negative and positive meanings of the well-known
pair lerrorise versus 'freedom fighter' suggest. Lexical analysis is therefore
the most obvious (and still fruitful) component in ideological discourse
analysis. Simply spelling out all implications of the words being used in a
specific discourse and context often provides a vast array of ideological
meanings. As a practical method, substitution of one word by others
immediately shows the different semantic and often the ideological 'effects'
of such a substitution.
Theoretically, this means that variation of lexical items (that is, lexical
style) is a major means of ideological expression in discourse. Depending on
any contextual factor (age, gender, 'racé, class, position, status, power,
social relation, and so on) language users may choose different words to talk
about things, people, actions or events. Personal and group opinions, that is,
attitudes and ideologies, of participants are a prominent contextual constraint, and hence a major source of lexical variation. Given the obvious
ideological implications of lexical choice, we may also expect that language
users are often (made) aware of their style, and may hence also partly
control it, and thereby either emphasize or precisely conceal their 'real'
ideological opinions. The current debate on 'politically correct' language,
precisely focuses on this aspect of ideologically based lexical style, and
especially shows people' s positions in the relationships between dominant
and dominated groups.
206
Discourse
Propositions
Beyond lexical semantics, the study of discourse meaning of course has
many other aspects that are relevant for the mapping of ideology on text and
talk. Thus, first of all, the propositions that represent the meaning of clauses
and sentences have an internal structure, of which for instance the various
semantic roles (agent, patient, object, etc.) may exhibit the ways participants
are associated with an event, actively or passively, responsibly or as
experiencers of events and actions. In other words, semantic structures result
from model structures. Such semantic representations are obviously a
function of how events are interpreted and evaluated (in a model), and may
therefore be ideologically controlled, depending on the group membership,
the position or the perspective of the speech participants. Who is seen as the
hero or the villain, the perpetrator or the victim of an act, which roles need
to be emphasized or concealed, are questions that organize many ideological
attitudes, and such perceptions may directly be mapped into propositional
structures and their variable syntactic formulations (actives, passives, nominalizations and so on). $
Local and global coherence
Whereas most of the structures mentioned aboye are within the traditional
realm of Iinguistic grammars, discourse analysis was precisely developed in
order to account for structures and strategies beyond the sentence boundary.
Semantics (as well as pragmatics and interaction analysis) is especially well
suited to account for such more complex 'textual' meanings. Thus, sequences of sentences (or rather, of propositions) constitute discourses if they
satisfy a number of coherence conditions, such as (a) conditional relations
between the 'facts' denoted by these sentences, or (b) functional relations
(such as generalization, specification, contrast) among propositions.
Such coherence is based on the interpretation of events as represented in
the mental models of the language users, and may therefore also be
ideologically influenced. Whether language users see a social event as a
cause or not of another social event may thus have an effect on the
coherence of their discourse. In other words, coherence is both contextually
and socially relative, and depends on our ideologically controlled interpretation of the world.
The same is true for the kind of overall coherence represented by topics or
semantic macrostructures, which also signal what speakers or recipients
think is the most important information of a discourse. Such a judgement
may obviously be ideologically based. What for some is defined, topically,
as a 'race riot by a violent black mob', for others may be semantically
summarized as an 'act of urban resistance against racist police officers'. In
other words, semantic macrostructures (derived by special semantic 'reductioñ rules or strategies from propositions in models about an event) not only
define such important discourse structures as topics, overall coherence, or
Discourse structures
207
importance of information, but essentially also explain the well-known
ideological practice of 'defining the situatioñ .
The implicit and the explicit
Another ideologically relevant property of meaning is propositional relations, such as implication, entailment and presupposition. Thus, information
that is explicitly asserted may emphasize negative properties of outgroups or
positive ones about ingroups, whereas the reverse is true for implied or
presupposed meanings. The well-known ideological function of concealing
'real' social or political facts or conditions may be semantically managed by
various ways of leaving information implicit. This also shows the importance of distinguishing between mental modeis (beliefs) and discourse
meanings, although we often may infer what people 'really mean' (their
modeis) when they say something.
Similarly, we may describe acts or events in great detall, or do so only
with few details, or at higher levels of abstraction. Such variation may also
encode ideological positions — who, indeed, has interests in knowing or
concealing such details about social events? In sum, semantics is a rich field
of ideological'work' in discourse, and virtually all meaning structures are
able to 'signify' social positions, group perspective and interests in the
description of events, people and actions.
Schematic structures
Whereas topics represent the global meaning of discourse, overall schematic
structures or superstructures represent the global forro of text and talk. Such
global discourse forms or schemata are organized by a number of conventional categories, such as introduction and conclusion, opening and closing,
problem and solution, premises and conclusion, and so on. Stories, news
reports, conversations, meetings and scholarly articles, among many other
genres, are thus organized by conventional schemas that define the order and
hierarchical position of such categories (as well as the semantic macrostructures or topics that define the 'content' of these categories).
As is the case for the syntax of sentences, also this 'discourse syntax',
may vary and hence'code for' ideological positions. As is true for all formal
discourse structures, these schemata may signal importance, relevance or
prominence. What information appears in a headline, what is emphasized in
a conclusion, or what event descriptions count as complication or a
resolution of a story, depends on the ways events are interpreted, and hence
on ideologically variable positions. Obviously, some of these categories are
obligatory (as is the case for headlines of news reports), but others are not
(for instance background information in news reports), and also categories
may appear in different positions. Thus, greetings and leave-taking are
usually obligatory categories of conversation. Besides interactional functions, for example of address and politeness, they may also have ideological
208
Discourse
functions, such as when their absence is intended as an ideologically based
insult Similarly, if verbal reactions in a news report appear up front, we
know that the source of such reactions is found important, as are his or her
opinions, a structural feature that obviously has ideological implications. 9
Rhetorical structures
Discourse features a number of special structures or strategies that have been
amply described already'in classical rhetoric, and that are usually called
'figures of stylé , but which will here be called rhetorical structures. These
structures appear at all levels of discourse described aboye, and assign
special organization (repetition, deletion, substitution, etc.) to these levels,
for instance by the figures of rhyme and alliteration at the level of sounds,
parallelism at the level of syntax, and comparison, metaphor, irony, etc. at
the level of meaning. Unlike other discourse structures, these are optional,
and serve especially in persuasive contexts, and more generally to attract or
manage the attention of recipients.
In an ideological analysis this will usually mean that rhetorical structures
are studied as mearas to emphasize or de-emphasize meanings as a function
of ideological opinions. Metaphors may be chosen that highlight the
negative character of our enemies, comparisons in order to mitigate the
blame of our own people, and irony to challenge the negative models of our
opponents. Rhetoric, defined in this sense, is essentially geared towards the
persuasive communication of preferred models of social events, and thus
manages how recipients will understand and especially how they will
evaluate such events, for instance as a function of the interests of the
participants. It is therefore not surprising that rhetorical structures play such
an important role in ideological manipulation.'
Speech acts
Whereas utterances were traditionally analysed along two main dimensions,
namely, expressions (signifiants) and meanings (signifiés), the philosophy of
language and the social sciences have added an important third dimension:
action. Uttering words and sentences in text and talk, in a specific situation,
is also arad at the same time the accomplishment of a large number of social
actions, as well as participating in social interaction. Thus, assertions,
promises or threats are made, arad such speech acts are typically defined in
terms of social conditions of participants, namely, their mutual beliefs,
wants, intentions, evaluations and goals that have social implications.
Speech acts like threats presuppose power, and tell recipients that the
speaker will do something negative if they do not comply with his or her
wishes. Commands also presuppose power but require that the recipient
must do something. That is, relations between speech participants are crucial
in the ways speech acts are accomplished.
Discourse structures
209
This also means that if these social relations are ideologically grounded,
for instance in relations of dominante and inequality, such relations may
well also be displayed in the kind of speech acts speakers are (or feel
themselves) entitled to accomplish. At this point, the ideological control of
social practices directly impinges on speech acts, for example when whites
of equal social position feel entitled to give orders to a black person, or when
men threaten women. In sum, whenever relations between participants as
well as other dimensions of the context (time, location, etc.) are ideologically based, this may show up in the kind of speech acts being
accomplished by the participants.
Interaction
Finally, within the vast field of the social actions being accomplished in or
by discourse, we fmd a number of interaction strategies that express,
indicate, reflect or construct specific social relations between participants,
and which therefore are ideologically relevant. It is especially at this level of
analysis that social position, power and control of social members may be
exercised, opposed, mitigated or emphasized.
Interactional control may affect virtually all levels and dimensions of text
and talk. Powerful speakers may control context structures by requiring or
prohibiting the presence of specific participants, setting a time or place,
allowing specific gentes and not others, prescribing or proscribing the
language or professional jargon spoken, by initiating or changing preferred
or dispreferred topics or an agenda for a meeting, by sanctioning formal or
informal lexical style, by being polite or impolite, by (requiring) the
accomplishment of specific speech acts or the management of tums at
speaking, or by opening or closing the interaction, among many other ways
text and talk may be controlled. In all these forms of control, it is the social
position of the participants, and more generally the ideologically based
interpretation of the context that is thus being enacted, expressed or
constructed in talk.
More specifically, the interaction dimension of discourse is relevant in
everyday conversation and other forms of spoken, face-to-face dialogues
such as meetings and parliamentary debates. Such conversations are organized by a number of specific structures and strategies, for example those of
turn-taking, interruption or beginning and ending. Many of these are
obligatory and hence not directly controllable by ideologically variable
contextual factors. However, as is the case for interaction in general,
ideologically based group membership, power, positive self-presentation or
outgroup derogation are among the underlying social relationships that may
impinge on conversational structures and moves. That is, who may (or must)
begin or end the conversation or meeting, who may initiate or change topics
or who may interrupt whom, are among the many forms of power display in
discourse that may also have an ideological dimension, for example those
based on gender,'racé or class."
210
Discourse
Ideology and discourse control
In these and many other ways, thus, we see most concretely how relations of
dominance, conflict or competition between speech participants may implement and enact relations between groups. People not only engage in such
verbal social practices as individuals and as cultural members, but also as
members of specific groups, and such identities and membership may also
be locally negotiated. That is, group dominance is not simply mapped on
contextual relations between participants, but may be flexibly managed and
exercised in situationally variable ways.
The same is true for the ideologies that sustain such practices. From the
abstract level of group representations, they may provide particular opinions
about other group members which together with specific contextual constraints provide the unique interactional configurations we observe in
ongoing discourse. More generally, also for the levels we introduced aboye,
ideological mapping on discourse structures is seldom direct. It takes place
through more specific group knowledge and attitudes, the formation of
'biased' models of events and contexts, the construction of meaning representations, and the expression in variable forms and surface structures, in
ways that are a function of many social and contextuál constraints, of which
ideological beliefs are only one element.
For the practice of ideological analysis this also means that ideologies
cannot simply be 'read off text and talk. What is an ideologically relevant
expression in one discourse or context may not be one in another, or may
have an opposed ideological function at another moment. This means that
ideological discourse analysis is very complex, and needs to take into
account all levels of text and context, as well as the broader social
background of discourse and interaction. In the following chapters, I shall
discuss some of the topics of such an ideological discourse analysis.
22
Context
What is context?
A broad characterization of discourse as a communicative event not only
features the various levels, structures or strategies of text and talk discussed
in the previous chapter, but also those of the context. Despite many informal
discussions in socio-linguistics, pragmatics and discourse studies of this
notion of context, there is strictly speaking no theory of what exactly a
'context' is. 1 The terco itself suggests that it is afl that comes 'with the text',
that is, the properties of the 'environment' of discourse.
I shall stay as close as possible with this linguistic version of the
commonsense notion of context, and define it as the structured set of all
properties of a social situation that are possibly relevant for the production,
structures, interpretation and functions of text and talk.
Thus, it is well known that, for example, the setting and the various group
memberships and positions of participants (e.g. age, gender, power) play a
prominent role in the way discourses are shaped and understood, and how
they function in the social situation. Other features of the social situation
may well be socially relevant but neither usually nor systematically influence specific structures of discourse, for instance the beauty, height or
clothing of the participants, although there may be some societies, cultures,
or situations where also such properties of a social situation become
contextually relevant for discourse.
Why is such a theory of context relevant for the theory of ideology? As
we shall see, contexts — defined as structures of discourse-relevant properties
of social situations — instantiate many properties of social events and social
groups that are monitored by ideologies. Thus, group domination, conflict
and competition will be multiply exhibited in everyday practices of social
actors, including their communicative practices. That is, ideologically relevant interests such as group identity, activities and goals, norms and intergroup relations of dominance and resistance, as well as social resources, are
also locally exhibited and reproduced in social situations, and hence in
communicative contexts. More speciflcally, we will find that ideologically
based dominance also involves the control of context. Specifying contexts
thus provides insight into the details of the exercise of social dominance and
its underlying ideologies.
212
Discourse
Context models
In most studies of context, typically so in conversation analysis, sociolinguistics, pragmatics or the ethnography of speaking, contextual properties
are assumed to directly affect (or be affected by) discourse properties.
Within the socio-cognitive framework presented here, no such direct relation
exists. Rather, the notion of relevance implies that models are relevant only
for language users, and hence only may influence discourse through the
ways they are being subjectively constructed by language users.
Such constructions again imply mental modelling. That is, it is not the
context itself (whether or not it 'exists' objectively) that influences text and
talk, but rather the context models of language users. 2 Such context models
are stored in episodic memory, just like the event models that are used to
represent what a discourse is about. Context models, thus, represent how
participants in a communicative event see, interpret and mentally represent
the properties of the social situation that are now relevant for them. This is
important, because it is precisely this subjective nature of context models
that also allows for personal variation and contextual uniqueness — it is not
the objective fact that speakers are women or men, white or black, young or
old, powerful or not, but how they see and construct themselves, in general
or in the current social situation. In other words, the essential pragmatic
notion of relevance may now simply be defined in tercos of context
models.
Context models are organized by the usual schema for interaction in
general, and thus feature a hierarchical structure of categories of the social
situation that language users find relevant for their production or reception of
text and talle. Below, I shall briefly discuss some of these categories. In a
later chapter, I shall explain how context models provide the 'personal'
interface between socially shared representations such as knowledge, attitudes and beliefs, on the one hand, and discourse structures on the other. For
now, it should suffice to say that context models monitor virtually all
'pragmatic' aspects of discourse, that is, all properties that may vary as a
function of the (interpretation of the) social situation, such as conversational
and speech acts being accomplished, as well as style, rhetoric and the ways
in which meaning incorporates information of event models (what people
know about an event being talked about).
As is the case for all mental models, also context models feature an
important evaluative component. That is, they not only represent the
knowledge or beliefs of language users about the social situation, but also
their opinions about it. Thus we may know our interlocutor, or the author of
an article in the newspaper, but also have an opinion about her or him, and
this opinion will of course also influence our interpretation of the discourse
itself, for example as more or less truthful or reliable. Similarly, our model
of the recipient (part of the context model) will also influence what we say to
him or her, and especially also how we do so, for example more or less
formally, intimately, politely or authoritatively.
Context
213
Obviously, as is the case for event models, such opinions may be
instantiations of socially shared attitudes, for example when men speak to
women or whites to blacks. Similarly, our beliefs about the current social
situation will also be instantiations of more general knowledge we share
with others about such situations, for example when we visit a doctor or
participate in a lesson at school. In sum, context models are also part of the
interface between socially shared representations and personal talk and text.
In part they simply represent what social members share, as well as their
own personal knowledge and opinions as based on their personal experiences, such as the beliefs about their friends, the settings of their everyday
conversations, their relevant goals, and so on. It is this combined personal
and social nature that makes models the necessary interface between social
cognition (and social structure) and discourse, between social macrostructure
and microstructure, and hence between ideology and discourse. Without the
notion of context model it would be impossible to explain how ideologies
may impinge not only on what we say (via event models), but also on how
we do so.
It should be emphasized that context models are not static but dynamic.
They represent the ongoing interpretation of language users of the social
situation. That is, context models may be partly planned, but ongoing
interaction and discourse, as well as other changing aspects of the social
situation, need continual updating of the context model. Indeed, during a
conversation or during the reading of a text, we may completely change our
initial interpretation of the genre (This is an interrogation and not an
informal chat') the goals of the speaker or writer this a threat or a
promise? ^ , and so on. This dynamic nature of context models also implies
that current fragments of discourse will become part of the 'previous
context' as soon as they have been accomplished.
This cognitive account of indirect, mentally mediated, context—discourse
relations does not mean that we reduce social contexts to cognition.
Obviously, contexts need their own social analysis, and so do discourses, as
forros of social interaction that are part of or constitute such contexts. It is
only the relationship between social context and action, on the one hand, and
the subjective understanding of context and discourse, on the other hand,
that needs such a cognitive interface. Indeed, without variable context
models, all language users of the same group would speak in the same way
in the same social situation.
As we have seen in Chapter 7, context models are a special case of models
of experience, which we construe from the moment we wake up in the
morning until we go to sleep at night: breakfast, going to work, and so on.
That is, the way we represent the social situation in which we engage in a
communicative event is only one of such daily experiences. Peoplé s
episodic or autobiographical memories are constituted by such models of
their personal experiences. Events of talk and text are merely special cases
of such mental models, namely, those that involve discourse as the relevant
event or action category. We still have few ideas about what such general
214
Discourse
experience models look like, but we may as sume that self occupies a very
central participant role in them. The subjectivity, perspective, point of view
or social position of self thus becomes the core of the model, which indeed
represents what I am doing when communicating.
Note, finally, that general experience models as well as context models
only represent specific, particular, concrete personal experiences of social
events. Episodic memory, however, also features more general information,
beliefs and opinions about oneself and others. Such general but personal
knowledge does not have the same episodic structure as models of events,
but is represented in a more abstract form. Yet, besides the influence of
socio-cultural knowledge and beliefs, also these personal beliefs are crucial
in the formation and updating of context models. In other words, in their
ongoing construction, context models are constructed from information from
the following sources: (1) a general schema, or goals or expectations about
the current social situation; (2) activated previous models (being reminded
of a previous conversation with X, reading the same newspaper in the same
situation, etc.); (3) general personal beliefs about such a situation ('My
neighbour always talks about bis work, and 1 don t like that'); (4) sociocultural knowledge and beliefs about communicative events (how to write
news stories, etc.); (5) previous parts of the ongoing discourse; and (6)
previous parts of the text. It is important to remember this variety of
underlying sources for context models, especially when we want to explain
how context models may be a function of social ideologies.
Dimensions of context
Let us now examine some of the situational properties that usually are
assumed to constitute the context, keeping in mind that it is not the
properties themselves that influence discourse (or that are influenced by
discourse) but their mental construction, as categories, in model schemata of
such social situations. In other words, despite the general, social and cultural
dimension of situational relevance, it is the personal construction of such
relevance criteria that for each discourse exercises the actual constraint on
current text and talk. Obviously, this also means that context models of
speakers or writers may be at variance with those of recipients, and lead to
communication conflicts about the 'definition of the current situation, as
well as with that of the group or culture as a whole.
Provisionally, then, we assume that the following situational parameters
may constitute categories of context models.
Domain
Communicative events are usually tied to a specific social or institutional
domain. In some cases, they may be constitutive of such a domain Thus, the
Context
215
many types of legal discourse constitute the domain of the 'law, whereas
types of political discourse largely constitute what we understand by
'politics' or the 'polity', and educational discourse the domain of education.
That is, a domain is the typical contextual property that defines overall
classes of genres, such as political discourse, medical discourse and scholarly discourse. For the definition of context they signal what social field the
context is a constituent part of.
For participants, contextual knowledge about domains serves as a global
orientation for the management of functions and circumstances of communicative events, for example in the use of professional jargon. Since domains
may be related to for instance professional group activities (e.g. those of
journalists in the domain of the media), and ideologies may be associated
with such groups, domains at the same time may function as 'ideological
domains', that is, as those sectors of society in which they define their
identity, exercise their activities, realize their goals, interact with relevant
groups and enact their power and where they protect or control their
resources.
In sum, ideological domains are sites of domination, struggle, conflict and
interests. Domains may be ideologically protected by groups as 'their'
domain, in which other groups should not 'interfere'. Thus, it is a main tenet
of market ideologies that the state should not interfere in the markets, of
joumalistic ideologies that the state should not abridge the freedom of the
press, and of professors that nobody should interfere with the freedom of
teaching and research. Many of the properties of discourse signal such
ideological embedding in social domains. Indeed, legitimation is a domainsensitive function of communicative events.
Overall interaction and type of speech event
For the planning, ongoing interactional management, understanding and
recall of speech events, participants need to be able to categorize them at an
overall level. They often use a name or genre description to do so. Thus,
participants may describe what they participated in as a conversation, a chat,
a meeting, a lesson, a parliamentary debate, seeing the doctor, reading the
newspaper, or writing an application letter, among a very large number of
other genres. These genres may then be characterized by several of the
discourse structures discussed in the previous chapter and by the context
features listed below. That is, genres are types of discourse that require
definition in tercos of both text/talk and context. Thus, genre knowledge of
participants will monitor many formal properties of discourse (such as
schematic organization and style) and well as the choice of topics. 4 If
ideologies are typically being reproduced in, for example, lessons, propaganda and news reports, then this reproduction process needs to be studied
for all relevant properties of the context as well as of text or talk itself of
these genres.
216
Discourse
Functions
The genres defined by the various context properties discussed here usually
have specific functions in an action sequence or domain, for example as
condition, consequence, purpose or goal of other social acts or events.
Exams for instance function as a test of the successfulness of educational
instruction and the qualification of students as participants; interrogations
are carried out with the goal of obtaining knowledge, typically about
criminal acts; parliamentary debates are constitutive of political decision
making; news reports are written and read in order to provide or to obtain
information and opinions about new events, and so on. In the accomplishment of their discourses, language users orient to these overall social or
institutional functions of the communicative event, and thus will adapt many
properties of their text or talk (or their understandings of such text and talk)
to these functions, either because such is the norm or rule, or because such is
strategically more efficient or successful.
Social actors as group members may of course have ideological representations of the functions of their discursive practices. Thus, journalists
may see their newswriting as serving as a 'watchdog of society', professors
their research as 'establishing the truth', and judges their judgements as
'doing justicé . Similarly, genres may have illegitimate or immoral ideological functions in the exercise of power, as may be the case for torture sessions
in order to obtain confessions, some police interrogations or racist propaganda in order to incite racial hatred. Well-known ideological functions of
discourse, to be discussed in more detail later, are, for example, those of
legitimation, defence and control.
Intention
Communicative acts, like all forms of action, are intentional. Theoretically
this means that participants construe mental models of what they want to do
(say, write) in the present context. The discourse itself is thus accomplished
in order to realize the intention and its represented outcomes. Obviously, and
especially in spontaneous conversation, such intentions may be negotiated
and interactionally modified or abandoned in the ongoing context. Despite
such possible modifications, speakers usually manage and execute their talk
and text according to their intentions, and often display such intentions in
various strategic positions of the discourse, for example by such expressions
as 'What I am calling for is .. 'What I wanted to talk to you about was
...', and'This article will be about.. . Ideologically relevant is precisely
the concealment of the 'real intentions' of speakers, for example when a talk
is announced as a friendly chat, but in fact is intended as an interrogation, as
political manipulation or as racist propaganda. That is, overall genre or
communicative action type are essentially related, from the point of view of
the speaker or writer, with specific intentions, but recipiente may or may not
be able to detect such intentions, and may thus be manipulated. I shall later
Context
217
discuss some of the mechanisms of such forros of ideological manipulation.
Several directions of research, and especially also conversation analysis,
reject analyses in tercos of 'intentions', for example with the argument that
such would ignore the fact that intentions (if any) are personal and private,
and only become socially relevant when displayed in text and talk. I disagree
with this position. First, intentions do play a fundamental role in social
contexts, namely, as necessary antecedents of social action, and there is no
reason to privilege the interpretations of recipients over those of speakers or
writers in that respect. Second, recipients continually construct the possible
intentions of speakers even when these are not fully displayed in text or talk.
From their own experiences (and their own intentions) they know that
speakers often do not say (exactly) what they intend to say, and recipients
may worry about that, ask about it, and otherwise topicalize intentions.
Indeed, they also think about them without saying so, and such thoughts may
again monitor what they will (not) say next. Third, for the same reason,
intentions not only monitor what is said or socially displayed, but also what
is not said. That is, the non-said may be interpreted by recipients also as part
of what speakers or writers intend with their utterance. Fourth, it is
inconsistent to reject intentions with the argument that they are personal or
private and relevant only when displayed in talk, and not do the same with
the (social?) understandings of the recipient. That is, hearer perspective is
not more social than speaker perspective, and both intentions and interpretations are both mental and social in interaction. Fifth, rejecting intennous as irrelevant is inconsistent with the broad acceptance of other mental
representations as underlying discourse and discourse production, namely,
knowledge and roles. In sum, a fully fiedged theory of discourse and context
1 irnpossible without assuming the relevante of intentions of speakers or
writers as part of the 'cognitive' dimension of the context. No doubt they are
crucially important for language users thernselves, given their frequent
references to, and inferences from their mutual representations of each
others' intentions.
There is no way to account for actions, and hence for discursive acts,
without their cognitive counterpart, namely, intentions as represented in
action models. These are integrated in the more complex model of the whole
communicative context. Theoretically, actions are combinations of such
intentions and the 'doings' that actually realize them, much in the same way
as discourse is a combination of (mental) meanings and the 'observable'
utterances that realize such meanings. Conversely, understanding an action
means the tentative reconstruction of an'intended' model, as inferred from
observable 'doings' in some context of interaction — co-participants or
observers try to figure out what actors 'mean' or 'intend' by their displayed
'doings'. We see again how closely discourse and (inter)action are related to
cognitive representations.
Intentions may appear purely individual and tied to the personal circumstances and biography of speakers. So, how could such a context category
218
Discourse
possibly have ideological functions? If ideology is defined as a group selfschema consisting of a number of categories, it soon becomes apparent how
intentions (as plans for action) may of course also be ideologically based.
They may represent (plans for) typical group activities (e.g. writing news
reports by journalists), include norms and values about how to do so
appropriately, identify the social position of self (as speaker) with that of the
group (e.g. as being dominant or not), and the implementation of specific
social resources (such as knowledge or access to public discourse). More
specifically, they may instantiate social attitudes, for example when whites
intend to derogate blacks with a s1ur, or when anti-abortionists intend to
argue against abortion. Indeed, as we see, speech acts as well as many other
discourse properties, as suggested in the previous chapter, may be a function
of ideologies, and this will often be the case intentionally for the speaker, or
as intentions attributed to the speaker by recipients in their model of the
context. In other words, ideologies often can 'reach' discourse structures
precisely through the intentions of the speakers: discourse is action, and
hence intentional, and such intentions may also extend to specific properties
of discourses. This does not imply that all ideological discourse structures
are intentional, or that ideological functions of discourse are always
intended. Speech acts by definition do, and so may topics and sorne elements
of style. More detailed surface structures (e.g. intonation or stress, syntactic
clause structure) or semantic dimensions may be sometimes intended, and
sometimes be more or less 'automatic' expressions of the representation of
the context. In other words, not all characteristics of context need to 'pass'
through the intention category.
Purpose
Often confused or collapsed with the notion of 'intention', also purposes
need to be introduced as a separate category in context structures. Thus,
whereas intentions are mental models of (discursive) acts, purposes or goals
are mental models of the broader consequences of such acts, for instance of
thefunctions of discursive acts as discussed aboye. Thus, contributions to a
parliamentary debate may have as their purpose to enact or defeat a bill, and
a lesson to teach students sorne knowledge or skill, as discussed for
discourse functions aboye. The difference between purposes (as well as
intentions) and functions, as defined here, is that functions are social, and
intentions and purposes are mental representations of speech participants.
This theoretical distinction is crucial. It allows us to assign different social
functions to discourses accomplished with the same 'purpose in mind', to
account for 'unintended' social consequences of intentions and purposes, to
describe and explain the role of individual speech events in the social
structure, to explain conflicts between purposes and functions, and so on.
Obviously, there are many forms of ideologically based implications of such
a distinction. Manipulation, for instance, is precisely to successfully accom-
Context
219
plish a speech event of which the recipiente do not know or understand the
ultimate purposes.
Date, time
Discursive events by definition have beginnings and ends. That is, they take
place in time, on specific days and dates, and for a specific more or less strict
or variable duration. Most official and institutional discourses (meetings,
appointments, sessions, etc.) have pre-set times of beginning and often also
of closing. Lessons and formal exams may last for instance one or two
hours. Sermons, depending on the religion, are usually pronounced on holy
days, and so on. Even informal talk has negotiated beginnings and closings,
namely, when people will meet, call or stop talking Newspaper stories have
deadlines and datelines, and people may read the newspaper only in the
morning or evening. Most informal and virtually all formal talk involves
ongoing time management. Speaking turns may be restricted or cut off by
chairs 'when speaker time is up', for instance in meetings, court sessions or
parliamentary debates. In relationships of dominante and inequality, for
instance, people may not be allocated speaking time at all, or their rightful
time to speak may be cut off. Such unequal treatment may be based on age,
gender, race, class, education, or status, and hence be ideologically based as
well as thus reproduced.
Location
Many communicative events typically take place in specific locations.
Where everyday informal conversations may occur virtually everywhere
(although in some situations they may be prohibited, e.g. during many
institutional communicative events: lessons, court sessions, meetings, etc.), a
lesson will typically take place in a classroom; an interrogation in a police
station or in court; a verdict always in court; a parliamentary debate in the
'Housé : and so on. Those in power not only set time and period, but often
also location for talk, as is the case for appointments of patients with doctors
or students with professors. Depending on power relations, similarly, journalists will go and interview important people where these want it and not
where the journalists propose. Hence, place of talk may be an element of
power, and hence ideologically relevant in the accomplishment of discursive
practices when location decisions break norms or mies of acceptable
communication as a forro of power abuse.
Circumstances
Many speech events can only take place when specific social or other
circumstances or conditions are in effect. A verdict can be pronounced only
'when the court is in session'; some meetings only when a specified number
of participants (a quorum) is present. These circumstances may themselves
be discursive, thus defining intertextual complexes like court sessions or
conferences. Sentences may be pronounced only after a verdict, and a
220
Discourse
verdict after indictments, defences and pleas, among other legal discourse
genres. Discourse may take place in 'inappropriate' circumstances, or
conversely, it may be inappropriate in the given circumstances. Such
communicative and social conflicts may play a role in the reproduction of
dominance: defendants may be interrogated without the presence of their
lawyers, women be made sexual advances at work, and so on.
Props and relevant objects
It may seem strange to include various props as part of a broader discourse
analysis, but if the analysis of context is part of such an extended account of
text and talk, then it makes sense to take them into account. Thus, the
context of a lesson may feature educational props such as a blackboard,
chalk, or an overhead projector, among other props, and usually relevant
furniture, of which the table of the teacher, usually placed in front, will be
different from that of the students. Similarly, a doctor's consultation may
have its own typical props, beginning with a white coat (at least in hospital),
a stethoscope, and many more, afi objects that are both indexical of ongoing
medical routines of investigation, as well as symbols of the doctor' s status
and role. Attorneys for the state or the defence may be asked to 'approach
the bench' by the judge, and such will also influence their manner of
speaking (confidential, non-public, whispered talk). Judges and chairs of
meetings will probably have and handle a gavel to open and close meetings
or to mark decisions being made, and so on for many other institutional
communicative events.
As suggested, these props may also be indicative of ideologically relevant
properties of the interaction or the social domain, such as hierarchical
relations and dominance Those who control meetings and sessions will
often sit in front, and possibly somewhat higher than the rest (as the judge
behind the lench', or the speaker of the house in a special position and
seat), if only to mark their current (powerful) role. Participants (police
officers, military, doctors, nurses, lawyers and judges, etc.) may wear
uniforms that index their position, profession, role or status, and these and
other props may be legally obligatory, such that the communicative event
may not even 'coune as a socially or legally binding act without them, or
they may be optional (as the flag in the president's or governor' s office) and
merely symbolic of the participants 'office'. From these few examples, we
see that props such as furniture, uniforms, objects, and so on, have many
social and symbolic, and hence ideological implications, and as such they
may also be represented in the context models of the participants.
Participant rolé
Social actors participate in communicative events in several types of role.
First of all they usually take part as speakers, writers, listeners (hearers) or
readers, and some of these roles, as is the case in interaction, will alternate,
such as that of speaker and recipient. But there are complications. The reader
Context
221
of the news on TV may well have a speaker role, but is not always the
person who actually wrote the news reports, say a writer, editor, or reponer,
who may have the role of the 'originatof of the relevant news text. Indeed,
we know that many other people may be involved in the joint production of
that news broadcast, such as producers, directors, camera people, reporters
'on location', and so on. That is, the production of institutional discourse
may have several layers or stages of actual text making, of which the person
who broadcasts, publishes or distributes may only be the last one.
The same is true for the various recipient roles. In a conversation with one
other person, this is easy and straightforward: the addressee is the same as
the hearer. But as soon as more people are present, they may not be there as
addressee even if they listen and hear what is being said; they may be there
in the role of 'overhearers' or as an audience in a talk show. And a talk show
will have participants who speak to each other, but the TV viewers are the
'real' addressees and listeners of this mass mediated dialogue. A doctor may
speak to trainees or nurses at the bed of a patient in hospital, but obliquely
also addressing the patient, or when speaking in medical jargon trying to
conceal from the patient what is being said. A secretary who takes the
minutes of a meeting is certainly supposed to hear what is being said, but at
the same time is seldom an addressee of the various turns at talk. A
defendant in court or during an interrogation at the police station may have
to talk when asked to do so (and when a lawyer is present), but may also
decide to remain silent under special conditions. Students in class are
expected if not required to listen and to speak when being asked to do so.
In sum, in most institutional situations there is a complex structure of
participant roles, usually defined in terms of the social roles of the social
interaction but in this case only defined in relation to the kind of contribution
they make to the whole event, what rights and obligations they have, and
hence who must speak or may speak, who must listen or may listen in a
given situation. Ideological dimensions of these various communicative
roles are as obvious as they are in social situations and practices in general.
Again, power and dominance may be enacted, and these may express
ideologically based inequality, for example when those in power abuse their
communicative roles, and prevent others from assuming their rights as
speakers or listeners, or force them to speak when they have the right to
remain silent.
Professional role
Similar remarks hold for the various professional roles participants exercise
when participating in communicative events. 6 The earlier examples also put
many of these on stage: professors, judges, police officers and so on. That is,
in this case the participants derive their communicative roles (e.g. as
speakers, producers, or chairs) from their socially or legally established
professional roles. Each of these professional roles may be associated with a
set of participant roles, as well as with types of communicative events or
222
Discourse
discourse genres. Thus, whereas public prosecutors have access to the genre
of indictments, and may interrogate defendants and witnesses, they are
obviously barred from verdicts or sentences, genres to which only juries and
judges have access, respectively. Similarly, professors have active access to
explaining things in a class and to asking questions during exams, and
students will have the obligation to act as respondents.
As is the case for the other•relevant categories discussed aboye, also this
category is a proper element of the context if it systematically relates to the
structures of text and talk. Thus, in their participant role as speakers and
their legal professional role as 'chairs' of trials, judges may control the kinds
of speech act (defendants must make assertions), topics (defendants must
speak about the facts being discussed), style (defendants must speak politely
or else may be held in 'contempt of court'), genre (defendants may or may
not be allowed to teli a personal story of their experiences), and especially
the many interactional features of the dialogue: defendants are not allowed
to interrupt the judge, they must begin and end their contribution when
required to do so, and follow many other rules of judge—defendant discourse
in the courtroom. Indeed, the whole point of a contextual analysis is
precisely to single out those properties of the communicative event that may
have such systematic relations with such properties of talk and text.
The ideological implications of these relations between professional roles
on the one hand, and participant roles and genres or speech acts on the other
hand, are fairly straightforward, as discussed aboye. As soon as professionals break the rules of communicative interaction and limit the rights of coparticipants they may enact forms of domination that may be based on
ideological beliefs. This may be the case for male doctors in relation to
patients, for male professors in relation to female students, and so on. Note
that here, as well as in the other examples given aboye, domination as well
as the ideology on which it rests, are enacted and thus reproduced by talk
itself.
Social role
In the complex network of various types of role of speech participants, we
may distinguish yet another type, which we may simply call social role.
Unlike communicative roles, these are not limited to contributions to text
and talk, and unlike professional roles, they need not be related to organizations and institutions. Indeed, these social roles obtain in virtually all action
and interaction. Por instance, whatever our position or professional role, we
may act and speak as a friend, an enemy, an ally, a proponent or an opponent
of other participants. Speakers in parliament, in their professional role of
parliamentarians (or cabinet ministers) affiliated with the legislature of a
country or state, may also speak as opponents of a bill, or as alijes of those
who introduced the bill. The same may be true in everyday conversations as
well as in formal talk. Such social roles will usually be enacted by specific
discourse features, such as forms of address, politeness moves, strategies of
Context
223
positive self-presentation (face-keeping) or negative other-presentation,
argumenta (against opponents, or in support of allies, etc.) and favourable or
unfavourable rhetoric.
Note that even within this category there are levels or layers of roles. In
everyday conversations or parliamentary debates opponents may be 'direct'
and confront each other face to face. In a debate on the op-ed page,
opponents may also be confronting each other personally, but not face-toface and not at the same time.' But, as is the case for various communicative
roles, there may be indirect, long-term addressees or relations. Opposing a
speaker may stand for opposing her or bis boss or organization, and speaking
as a member of an action group may be interpreted as advocating the stance
of the action group itself. Tbus, in parliament, speakers may oppose what the
previous speaker has said, they may, more broadly, oppose the bill being
proposed by someone, and by so doing they may oppose the party to which
that person belongs, and at the same time they of course advocate their own
position, and/or that of their own party (which need not be identical), and as
political representatives they may at the same time represent or oppose the
'special interests' outside of parliament. In other words, deeper and more
sophisticated analyses of contexts in principie uncover complex sets or
levels of various roles.
We have seen that besides membership of groups and organizations,
ideologies typically involve polarization, struggle, conflict or competition,
and these relationships precisely map onto the social roles being introduced
into the context here. Ingroups and outgroups and their associated ideologies
thus manifest and reproduce themselves precisely by the 'position' their
members take in situations of debate and conflict, also in communication.
Arguing in favour of a bill that restricts immigration, may by its very stance
be part of the reproduction of nationalism or ethnocentrism. An anticommunist speech in parliament is thus taking a stance in an ideological
conflict. In other words, social roles are contextually variable enactments of
positions, including ideological positions.
Affiliation
Participants in professional roles often doñ t speak 'for themselves, but as
representatives of an organization or institution, and as representatives who
in principie can be replaced by any other institutional member. That is, their
affiliation plays a prominent role in the context: confessing to a police
officer or in court, doing an exam, making a declaration for a tax auditor,
and so on, are the kind of speech events that are often appropriately
accomplished in the presence of any representative (in the same professional
role) of the organization. People speak in parliament or congress, but usually
do so as representatives of their partes, as they do when listening to such
speeches. More generally, these events and their participants are also
integrated in a web of institutional affiliations. Some of these may be very
224
Discourse
strict, and legally well described (also with respect to kinds of communicative events), while others may be looser and open to variation and
negotiation. Thus, teachers will usually have more leeway in the accomplishment of their communicative and professional roles than speakers of
parliament orjudges.
One of the many implications of the institutional or organizational
affiliation of communicative events is precisely the fact that such participants take part as representativas of the institution, and hence often carry the
institutional ideologies, if any, into the ongoing context. Indeed, the representatives of an organization are by habit, norm or law entitled or obliged to
represent the Interests' of the organization, and hence their talk and text will
multiply index or signal such ideological comtnitments. Thus, a teacher may
thus implement the educational ideology of the school or university, the
joumalist an ideology of the presa, and so on. Such ideologies may pertain to
the content of text or talk (such as newsworthiness of events for news
interviews or news reports), but also to the very nature of the interaction
itself. Educational or medical ideologies may or may not allow a more or
less independent and autonomous initiative to students or patients, depending on whether the ideology is more authoritarian or permissive.
Membership
More generally, participants may speak, write, listen or read (also) as
membeis of groups or social categories, in addition to the organizational
affiliation and the various roles described aboye. People may be male or
female, white or black, old or young, and so on, and either they themselves
or their co-participants will categorize them as such, and act (speak, write)
accordingly. S ince such social groups and categories are the basis of
ideologies, these ideologies will in principie also exhibit in the relevant
communicative social practices in which group members engage. That group
membership affects the structures of text and talk themselves has been
shown in much socio-linguistic research, for example on intonation, lexical
items, topics, rhetoric or interactional moves, as discussed in the previous
chapter. In terms of the context, people of different social groups or
categories are defined and treated as such, also in the communicative event
— they may be given preference in tutti taking, freedom in topic selection or
style, but they may also directly be discriminated against along the same
fines, only because they are a member of a specific group. Probably more
than any other category of the context, thus, social group membership is
what projects ideologies into communicative events. Later we will see how
this is being accomplished in text and talk. 8
It should at this point be emphasized again that roles, affiliation and group
membership are not always 'given' in social situations, and this is a fortiori
the case in subjectively construed models of such social situations. That is,
such social 'positions' may be negotiated, changed, oriented to, deviated
from, ígnored, forgotten or otherwise become less (or more) relevant in a
Context
225
specific situation. That is, a dynamic theory of discourse emphasizes such
situational and personal fiexibility. The same will be true for the ideological
conditions and consequences of the ways such categories are constructed in
the current context by the participants. Men may temporarily disassocíate
from their group and speak on behalf of women; speakers may defend the
position of their opponents when they act as devil' s advocate; and dissidents
are by definition speaking in defiance of dominant group ideologies.
The social others
So far, the relevant participant roles discussed aboye pertain to people
involved in various capacities in the communicative context itself. However,
text and talk are often also about other people, usually people who are not
present in the' ongoing context at all. Strictly speaking this is a property of
the nteaning of discourse, and hence part of a semantic and not of a
(pragmatic) context analysis. That is, discourse referents are .not part of the
context model, but part of the event model (partly) expressed by the
discourse. Thus, men routinely speak about women, whites about blacks,
and doctors about patients, and these social others are thus the referents of
their talk. It is also in this way that the ideologies relating communicative
participants to the social others, as members of outgroups, are projected into
the meanings of a discourse. Yet, one might also argue that these social
others are some kind of 'absent participants' in the context.
Racist talk addressed to other whites may obliquely be addressed, in a
broader social context, to the social others, and thus not only be semantically
relevant, but also pragmatically, that is, as an inherent element of the act of
discriminatory talk, as a form of reproduction of racist ideologies. That is,
the social others, as part of the targeted outgroup, may be talked about but at
the same time indirectly, socially and ideologically addressed. That is, acts
of discrimination also may be categorized and interpreted as such when the
discriminated party is not present — yet, they are somehow 'party' to such
communicative interaction.
Social representations
Most context categories discussed aboye have a proper social nature, and are
typically made explicit in sociological tercos. Obviously, however, participants not only have positions, rights, duties and relations in social situations,
institutions and overall social structures, but also share social representations, such as knowledge, attitudes and ideologies. Some of these mental
dimensions of participants have been discussed in terms of intentions and
goals, which are more individual and contextual.
However, especially as members of various social groups and institutions,
communicative participants also share social representations that have an
impact on ongoing interaction, text and talk. Thus, crucial for all communicative events is the respective knowledge of the participants, both personal
as well as social and cultural. Thus, speakers have knowledge about each
226
Discourse
other (that is, they have a model about Self as well as about others), and such
knowledge may instantiate more general knowledge and beliefs about the
group to which the others belong.
Similarly, ideologies of participants in many ways affect the ongoing
definition of the communicative situation, the various actions, participant
roles, as well as the discourse itself. The same is true for the socially shared
attitudes monitored by these ideologies. Indeed, these attitudes may even be
specific and tailored to the communicative event at hand. Thus, trivially, in
an informal discussion about abortion, or a parliamentary debate about
nuclear energy, speakers bring to bear their specific actitudes about these
issues, and such attitudes multiply influence the event and context models
that monitor ongoing talk: who is defined as proponent or opponent, whether
a speaker is seen as a representative of a social group (man or woman in the
abortion debate), who .will be treated more or less politely, and so on.
In sum, all social aspects of the complex communicative event are
variously monitored by the social representations of the participants as
members of groups, categories or institutions. Knowledge will be mutually
presupposed accordingly, for example when doctors or lawyers speak with
members of the same professional group, or when women presuppose both
knowledge, attitudes and ideologies of other women of the same feminist
movement. Indeed, most of the communicative context and the discourse
need not be made explicit because of such presupposed sharing of social
representations within the same group, society or culture.
Together with mental models of individuals, social representations are
part of the cognitive interface between social structure, group membership
and discourse. If people speak or write as members of groups, their group
membership will largely be brought to bear in the current context in tercos of
the social representations shared with the group, that is, as instantiations of
group knowledge, attitudes and ideologies.
This does not mean, incidentally, that social representations, including
ideologies, cause or determine text and context. It has been explained in
some detall in Part I, that there is still a vast 'mental distance' between social
representations, and hence the influence of social groups, on the one hand,
and discourse structures (including context) on the other hand. Most
crucially, although variably so in different situations, speakers are also
individuals with their own biography, goals, preferences, plans and emotions
— that is, with their own personal models. Intentionally or unwittingly, such
models may instantiate shared elements of social representations, but even
then the context and the individual and hence their text or talk will be
unique. If not, and as suggested before, all members of a group would say or
write the same thing in the same situation. This is also one of the reasons
why I include relevant aspects of personal models (e.g. intentions and
purposes) in the current context.
Social representations may not only apply to the semantic dimension of
discourse (e.g. abortion as a topic of talk), but also to the discursive
interaction itself: who may/must speak/write about what/whom, to whom, in
Context
227
which way? Joumalists know how to interview news sources or news actors,
how to write news reports and follow rules and strategies they have leamed
as group members, and the same is true in afl other professional roles
discussed aboye. Thus, both in conversation and in parliament, people
instantiate the very ideological forms of membership that we routinely
assign to speakers: he is a conservative, she is a liberal, and so on.
At this point we have come full circle. Ideologies may indirectly control
the properties of all categories of context models for discourse. But it now
appears that one of these categories itself pertains to the social beliefs, and
hence the ideologies, of the participants. In other words, ideological control
is, so to speak, not external nor deterministic, but internal, that is, through
the beliefs of the participants themselves. Thus, I may participate in a
conversation as an anti-racist, and this stance influences the way I construct
the current context as well as what I say and how I say it. At the same time,
both the recipient and I myself represent (part of) my anti-racist beliefs as
part of our respective context models, of ourselves as well as about each
other (indeed, I may know that my interlocutor knows that I am an antiracist, and may shape my talk accordingly).
There may even be a discrepancy between my role and my role as
represented in my model of myself in the present context. People may speak
as anti-racists without much self-control or self-monitoring and thus more or
less directly express and enact their group membership. However, they may
also do so by monitoring their current identity and by carefully managing
their'imagé as an anti-racist, for instance for recipiente that are hostile to
anti-racists. Also this subtle interplay between 'real' social identities of
discourse participants, on the one hand, and those that are locally and
intersubjectively represented in their current context models and displayed in
their discourse, on the other hand, shows how complex the relations between
ideology and discourse may be.
Concluding remark
The context analysis presented aboye shows that the discursive reproduction
of ideologies also applies to the contextual aspects of communicative events.
Contexts, or rather context models, explain personal, situational and social
variations in the ways underlying ideologies may or may not affect text and
talk. They thus serve as another layer of constraints, another interface,
between ideology and discourse, and explain that ideologies are not 'deterministic' in the sense of necessarily affecting discourse structures — this will
always, literally, depend on the context. Therefore, no discursive theory of
ideological expression and reproduction can be adequate without a detailed
analysis of context. We shall later spell out in somewhat more detail how
exactly mental models of such contexts intervene between social representations, including ideologies, and structures of discourse.
23
Reproduction
What is reproduction?
It has often been argued, aboye, that ideologies are typically reproduced by
social practices, and especially by discourse. What exactly does this mean?
As with most general notions, the concept of reproduction is not very
precise. In general, it implies that ideologies are 'continued', 'made to
remain, last, persist.. .', and so on. Like its second part, however, it implies
an active, human dimension: It is what people do, make happen, while also
making something new, creating something. The repetitive'ré part implies
that the act of production is being repeated. For social practices and
discourse this usually implies that such acts of production take place every
day, are routine, and are part of the definition of everyday life.
More specifically, however, when we refer to the reproduction of ideologies, we are dealing with an equally vague sociological notion, also used to
denote the reproduction of groups, social structures, or even whole cultures.
Again, reproduction here implies continuity of a system or structure as well
as human agency. More theoretically, the notion is used to bridge the wellknown gap between the macro-level and the micro-level of social structure.
Systems or abstract structures, such as ideologies, natural languages, and
societal arrangements are thus said to be both manifested in, as well as made
to persist as such through, social practices of social actors at the micro-level.
A language Like English is reproduced, daily and by millions of people, by
its everyday use. And so are capitalist, sexist or racist ideologies.'
The active concept of 'production' is relevant here because such systems
are not only being 'applied','implemented' or passively 'used', but at the
same time constituted and reconstituted, as well as gradually changed, by
such contextual uses by many social actors. Indeed, also the gradual
development of ideologies of a group is based on such social practices. That
is, ideologies are (re)produced as well as (re)constructed by social practices.
There is another macro—micro dimension involved here. This time not just
that of an abstract system on the one hand and actual practices on the other
hand, but the distinction between the group and its members, and especially
its new members. Just as groups are reproduced (also) by getting or
recruiting new members, also ideologies are reproduced by getting new
'users,' as is also the case for natural languages. Whether by socialization or
other processes of sharing social representations (initiation, teaching, train-
Reproduction
229
ing, preaching, propaganda), ideologies are continually reproduced because
new social members 'acquire' or learn to use' them.
As we shall see in more detall later, this may happen directly through
explicit ideological discourse, or indirectly by making inferences from
discourse and other social practices about what opinions other group
members share. White people learn racism' by accepting general racist
statements such as 131ack women are welfare queens' as expressed in
conversations with friends or colleagues, or they infer such a belief from
repeated stories in the media in which black women are portrayed as being
on welfare, or because they overgeneralize from one or a few black women
they know who are on welfare. This last case, as a personal experience,
however, is usually told in stories to other group members, and the relevant
inference may then be jointly produced in talk, as a conclusion suggested or
accepted by co-participants. That is, sharing is usually not simply a onesided, passive event, but a complex, co-operative procedure, involving
people who (already) 'know', as well as people who 'still don' t know'. In
other words, reproduction also implies socialization, leaming, inculcation or
adoption by young or new members, of the socially shared representations of
a group.
And finally, besides its macro—micro (system—actions, group—members)
dimensions, we also have the local and contextual versus the global and
decontextualized dimension of reproduction. Members having learned how
to make an inference from one case or example, or to express an ideological
opinion in one context, are typically able to do so for similar cases and in
similar contexts. That is, reproduction is not only top-down and bottom-up,
but also allows for transition from token to type and from type to token,
from today to tomorrow, and from here to elsewhere. Reproduction thus also
implies generalization. Combined with the vertical relations between system
and actions, this also explains the bottom-up nature of reproduction — social
representations are not merely acquired directly, in an abstract (and usually
discursive) manner, but also as generalizations from daily experiences. In
specific social situations of ethnic inequality, such generalizations may be
morally unacceptable overgeneralizations (prejudice), but they may also be
forms of (correct, justified) social learning, for example when minorities
learn to detect and interpret racist events as such, and thus acquire an antiracist ideology. 2
Summarizing these various aspects of the social reproduction of ideologies, we thus have the following dimensions.
1 System—Action: top-down application, use and implementation of general, abstract ideological beliefs in concrete social practices.
2 Action—System: bottom-up sustaining, 'continuing and changing the
socially shared system by its daily uses in social practices. Along this
dimension, ideologies are effectively being constructed, constituted and
changed by social practices, including discourse.
230
Discourse
3 Group—Mernbers: ideological communication, inculcation, teaching,
socialization and initiation of new members by (knowledgeable) group
members.
4 Members—Group: acceptance and compliance or non-acceptance, resistance or dissidence of one or some groups members, against the ideology
of the group or its elites.
5 Local—Global: generalization, extension, decontextualization of specific
experiences and opinions to similar or abstract contexts, experiences,
cases or circumstances; social leaming, overgeneralization, stereotyping,
prejudice formation and ideology construction.
In 4 we see that the group—member relation may also be conversed, that
is, when individual members reject, refuse or do not accept a group's
ideology. This may not seem to be a dimension of reproduction, but it is
necessary to account for personal variations and change of ideologies, which
are also part of their reproduction. Obviously, as soon as most members
reject ideologies or some ideological beliefs, then change may eventually
lead to the abolition of ideologies. 3
Discourse and reproduction
Many of the types and modalities of reproduction discussed aboye appear to
be discursive. Ideologies may be expressed in many genres and contexts of
discourse and their respective structures as discussed in the previous
chapters. Such ideological discourses have several functions, such as a
display of group knowledge, membership and allegiance; comparison and
normalization of values and evaluation criteria; evaluating social practices;
socialization; or persuasion and manipulation. Some of these functions will
be dealt with more specifically in the next chapters. Here I focus on some of
the more general aspects of the discursive reproduction of ideologies.
Context
In the previous chapter we saw how ideologies may interven in the social
construction or interpretations of the contextual categories which in turn
constrain (or are influenced or constituted by) text and talk. Thus, participants may act as speakers, as proponents, as journalists, as representatives of
an institution like a newspaper, and as members of various groups (age,
gender, ethnicity, nationality, etc.). In all these roles, participants may enact
(and sometimes disregard) the social representations, including the ideologies, related to their social identity. That is, social situations in general, and
contexts of discourse in particular, are literally the site where ideologies are
being enacted in society. As long as speech participants identify with or
willingly or unwillingly (have to) represent the groups and institutions of
which they are members, they thus by definition contribute to the use and the
reproduction of the ideologies associated with these social formations.
Reproduction
231
The examples mentioned in the previous chapter suggest, however, that
such ideological alignment is not straightforward. First, language users may
have their own personal models, and these may be more or less at variance
with the social representations they share as group members, given the
constraints of the present context. Indeed, their interests as group members
may be less salient or less relevant than their current personal interests, and
their intentions and goals may be formed accordingly. Second, language
users are members of several social groups, and thus share in several social
representations at the same time. Again, some of these may be more relevant
or more powerful than others. The result is that the event and context models
that monitor the communicative event may have contents and structures that
in many ways are inconsistent with those expected of loyal group members.
If such is the case for models, this will also be the case for the discourse
properties that are a function of these models, such as the meaning derived
from event models (including specific opinions), as well as the surface
structure, style, speech acts or interactional strategies that are controlled by
context models.
The consequences of these complex and subtle acts of interactional and
communicative management in specific social situations are that ideologies
are not simply reproduced in talk and text by the members of the groups that
share such ideologies. There is more or less substantial variation, there is
explicit and intentional deviance, there are dilemmas, and there are personal
and interpersonal conflicts that need to be negotiated and resolved. 4 Hence,
not all news reports in a newspaper will show the ideology or political
allegiances of that newspaper. Not all journalists always give priority to
joumalistic ideologies in their reports, and not all racists will treat minorities
always and everywhere with derogatory remarks.
The empirical picture emerging from this variation may be that ideologies
do not seem to 'exise in the first place — the local and personal constraints of
context may distort or prohibit their unfettered expression. The question is
then in what respect we are able to speak of the 'reproduction' of ideologies,
when social situations so often prevent their direct implementation. Theoretically, then, we are able to account for ideological reproduction only when
we assume that across language users and contexts, there are 'enough'
instances of ideological expression.
How much is 'enough'? Obviously, this may vary. Por instance, it may be
assumed that journalists most of the time will have to follow the ideological
principies of their profession. If not, they will not be hired or they will be
fired. Exceptions will be allowed, especially for highly qualified or popular
journalists, but there will be a margin of variation within which each
journalist will have to remain when working for one of the mass media. In
some cases, for example in public office, even one deviation from the
ideological 'party liné may be enough for a politician to be marginalized,
discredited or voted out of office.
Interestingly, quantity as such may not be the right measure. One public
racist statement may be enough to conclude that someone is expressing a
232
Discourse
racist ideology, even when in most other situations these expressions were
better controlled. The rationale behind such a conclusion is that people who
do not have a racist ideology will simply never make such a blatantly racist
remark in the first place. In actual practice, there will be a broad range
between regular and unique expressions of ideology, on which basis other
participants and observers will be able to draw conclusions about the
underlying ideologies of group members. Some of these expressions may be
very indirect or subtle, and participants and observers may not even notice
them if the ideology that inspires them is taken for granted. Thus, the quality
press, including the liberal quality press, may not daily make blatant remarks
about ethnic minorities or immigrants. Yet, more subtly and indirectly, for
example by the choice of its topics (e.g. about crime, violence or cultural
deviation), it may well slowly create a negative image of the cultural others,
and thus contribute to the reproduction of an ethnocentric ideology.
Given the processes of memory, attention and recall, readers may selectively focus on and memorize even the occasional story in which minorities
are represented negatively, and forget about the larger number of negative
stories in which members of their own majority group are represented
negatively. This is a familiar finding in differential attribution for ingroups
and outgroups. 5
In sum, the conditions of reproduction are as complex as the structures of
context and discourse, and the strategies of information processing and
social representation, combined. Under what conditions specific text and talk
is being attended to, read or listened to, understood, and represented in
models, and under what conditions these models are accepted as true and
generalized to more abstract social knowledge and beliefs, are all questions
that need to be answered in a theory of reproduction.
All this also applies to the projection of ideologies in context models and
hence in the enactment or interpretation of the context itself. Negative
beliefs about minorities when uttered by prominent members of minority
groups themselves or by a white cabinet minister of a respectable party, may
be much more credible than those of a member of a racist party. That is,
credibility is one element of the process of acceptability, and itself a function
of the group membership of the speaker, that is, a category of the context.
Generally, thus, acceptability of beliefs, which is the core criterion in the
reproduction of ideologies, depends also on the interpretation and the
evaluation of context structures, and especially on the various roles and
positions of the participants. Even the context categories of communicative
domain, action type, and circumstances may be especially conducive to
ideological reproduction, as is the case for classroorns and education,
parliament and politics, newsrooms and the media. This is so first because of
the credibility or the prestige of the social actors involved, as well as the
mass-mediated consequences of text and talk. One 'unhappy' but widely
publicized remark of a prominent politician about immigrants may contribute more to the reproduction of ethnic prejudices and ideologies than
Reproduction
233
thousands of blatantly racist conversations taking place in the homes of
citizens.
Discourse structures
Although context categories themselves may strongly influence the acceptance of social representations, the really influential factors should generally
be searched in the discourse structures themselves. That is, are there
discourse structures that prohibit, impair or favour ideological reproduction?
The analysis of the structures and strategies of discourse in Chapter 21,
suggests that ideologies may in principie map onto all levels and dimensions
of discourse: graphics, intonation, syntax, local meanings and coherence,
topics, style, rhetoric, speech acts and interactional features. Still, expression
structures as such usually do not code for ideology — this mostly happens in
relation to underlying meanings and functions. To persuasively convey
ideological 'context', thus, the semantics of text and talk plays an especially
important role.
To prove such an assumption, we need to find out how semantic
variations have different consequences for the construction of models, and
how friese models may in turn be used to confirm or construct social
representations. For instance, topics or semantic macrostructures of discourse represent salient and important information, and will therefore
generally be attended to, and be used to construct key (top) propositions in
models. If such topics are repeated (e.g. 131ack West Indians rioted' in the
popular press in the UK), then model construction may become routine and
generalized to a negative attitude about black youth, or about blacks in
general, if no alternative, counter-ideologies are present that may cause
rejection of such models.
At the same time, readers with ambiguous attitudes about minorities, may
find such prominent expressions of bias too crude to be credible, and may
not construct the biased models as intended. They may, however, be unable
to detect more subtle forms of semantic ethnic bias in news reporting, and
following their interpretation construct models whose generalization also
leads to a negative attitude about minorities. That is, besides contextual
conditions of credibility, also the nature of the semantic (and other)
structures may (for different participants) have different influences on model
construction and acceptance, and on the subsequent generalization to social
representations that are part of ideological reproduction.
Reproduction, however, is not limited to interpretation and the influences
of discourse on mental representations. Also the production side of the
communicative event needs to be taken into account. Part of this has been
done in the analysis of context. This means, among other things, that access
to specific social roles, and especially elite roles, provides group members
with vastly more influential means to reproduce ideologies than ordinary
citizens without much access to public discourse. These, then, are the now
familiar social conditions that control the context of production.
234
Discourse
But besides these contextual categories of position, roles and group
membership, we also peed to establish which discourse structures can be
more or less explicitly controlled in the first place. Some of this control, as
is the case in TV programmes, may be the result of complex production
processes. Ideological control in that case presupposes that most participants, and at least the more influential ones, are ideologically on the same
liase. Another questions is whether speakers or writers who have control over
discourse are always able to 'translate' their ideologies into the more or less
subtle properties of text and talk.
Thus, again, the explicit choice of negative topics in order to derogate
outgroups, is fairly easy and straightforward, and simply involves the
projection of ideologically biased models of events onto topics of talk and
text, as is the case for crime stories about minorities. However, many other
discourse structures, such as the syntax of headlines, local semantic disclaimers, or the choice of metaphors, is only moderately or not at ah
consciously controlled. Ideological influence of discourse in this case is
barely intentional, but a more or less automatic expression of biased
models.
Of course, this does not prevent ideological reproduction. On the contrary,
since it is not consciously controlled, it cannot usually be 'self-censored'
either because of prevailing norms or values (e.g. those of nondiscrimination), so that ideological reproduction takes place without the
speakers' being aware of it. Indeed, when confronted with critical analyses
of such subtle racist practices, they will generally deny that they are racists.
Thus, besides explicit manipulation of models and social representations,
ideological reproduction may more indirectly and unintentionally take place
through the routine and taken-for-granted processes of discourse production.
In the chapters that follow, I shall study a number of more specific instances
of these various aspects of the discursive reproduction of ideologies.
24
From Cognition to Discourse
Introduction
After the more general outline of the role of discourse in the reproduction of
ideologies in the previous chapter, I am now in a position to detall some of
the componente of a relevant theory of discursive ideological reproduction. I
shall begin where I left off in Part I, at the cognitive level of analysis, and
then I shall move to the various structures and strategies of text and talk that
are relevant for the expression of ideologies.
It should be recalled here that the cognitive basis of a theory of
ideological reproduction is neither a luxury flor a reduction of the social to
the personal. First, I have stressed that the mirad is social — socially acquired,
sharecl, used and changed. Many aspects of social structure presuppose such
shared knowledge and beliefs of members. A large part of our mirad consists
of socially and culturally shared representations. These are also needed in
the understanding of personal experiences and the accomplishment of
individual actions, and hence also for discourse production and understanding.
Thus, second, if we want to describe and explain how group ideologies
affect discourse, and vice versa, we need to spell out how to get from social
representations to the individual ones that represent personal experiences or
personal text and talk. The only way to do this is in terms of a cognitive
theory of discourse processing.
There is at present no serious alternative theory that explains how social
structures, including those of communicative contexts, are able to constrain
the structures of text arad talk. We simply peed the theoretical construct of
people' s 'minds' as ara interface between the social arad the personal. As is
the case for all theories, however, these may change, so that the mental
'architecturé as it was adopted from current cognitive science is of course
merply a hypothesis about the ways people produce and understand discourse and accomplish many other tasks.
The same is true for the'information processing' metaphor prevailing in
cognitive science. This is at present the only viable theoretical framework to
account for language use, communication and the ways knowledge and other
(e.g. ideological) beliefs interact with discourse. However, it was also
emphasized that such a framework is incomplete when it is not embedded in
a broader theory of (verbal and other) social interaction and social structure.
That is, beliefs and discourse have both cognitive and social dimensions, and
236
Discourse
the crucial point of this book is precisely to connect these two major
dimensions.
Discourse production
Discourse production involves a set of representations and complex operations that together may be thought of, theoretically, as a discourse production unit in the rnind. This unit has three main modules — a pragmatic, a
semantic and a formulation module — which operate in Glose collaboration
The pragmatic module
As soon as people want to speak or write, they first construct a relevant
context model. This model selects the relevant information from the speaker's beliefs about the social situation, as described in Chapter 22, for
example the current communicative event (e.g. informal conversation with
friend, writing a news report, giving a lesson, or visiting oné s doctor),
current goals or intentions, a setting, and the speech participants. As suggested before, such a context model may simply be a relevant specification
of the current experience model speakers have of the ongoing episode. 2
The context model thus specifies what relevant speech acts must be
accomplished, and generally provides the information needed in the other
(semantic, formulation) modules for the production of a discourse that is
appropriate in the present context. In other words, a context model contains
a'plañ that features all information needed to accomplish an appropriate
speech act. For instance, beliefs about the nature of the social relation
between speaker and hearer provide the relevant information for the accomplishment of deference or politeness, such as specific pronouns or the use or
avoidance of specific lexical items. Ah possible variations of discourse
structures that are not a function of the semantic module are controlled by
the pragmatic module and its current context model. That ís, speech acts,
interaction, as well as the stylistic and much of the rhetorical dimensions of
text and taik, are controlled by this pragmatic module.
In other words, whereas the semantic module specifies what people want
to say or write, the pragmatic module controls how they must do so in an
interactionally and socially appropriate and effective way, how discourse
'fits' the current context, and what social acts are accomplished by the
discourse.
Whereas in writing or monological communication, the context model
may be relatively fixed during production, in conversational interaction such
a model is of course continually updated, according to the feedback received
from other participants. Models of each participant in a communicative
event will partly be identical or similar, but also partly different — each
participant interprets and representa the 'current context' in an at least
slightly different way. These different constructions may be the basis of
From cognition to discourse
237
corhmunicative misunderstandings and conflicts, although language users
have effective strategies to solve such problems of misunderstanding.
The semantic module
The semantic module provides the information needed for the meaning
construction of discourse. It may draw on virtually all representations in
personal and social memory. This is not surprising, sine we may speak
about virtually everything we know or believe, including what other people
know or believe. In order to talk about past and current personal experiences, as well as about intentions for future actions, or what language users
know from others or the media about any situation or event, they draw upon
relevant experience and event models in personal memory. But they know
and believe much more than the specific facts represented in their models
about personal experiences. For instance they may also want to express
social representations, namely, what we know and believe in our group or
culture.
Obviously, people do not usually express all they know or believe, simply
because all this would not be relevant in the present situation, because the
recipients already may know or believe many of these things, or because for
whatever reason they do not want the recipients to know what they know or
believe. These constraints are contextual and therefore provided by the
pragmatic module and the information in the context model (the representation of what the speaker believes about the beliefs of the recipient).
Generally, then, only a small fragment of contextually relevant information
of event models will be selected for the construction of discourse meaning.
Other information may be left implicit, and may at most be signalled by
appropriate discourse structures, so that the recipients will be able to infer it
when they need or want to do that. Obviously, the more beliefs already
shared by the participants, the more discourse may leave meaning (representing such beliefs) implicit.
The output of the combined (ongoing) operation of the pragmatic and
semantic modules is a semantic representation. Whereas our knowledge, as
represented in personal event models, may well be accessible and available,
we usually do not know in advance what model information will be included
in this semantic representation. That is, language users have recourse to
effective strategies that allow them to continuously adapt the selection of
what they know and believe to the constraints of the ongoingly constructed
and updated context model (e.g. what they think is interesting for the
recipients, what they need to say in order to remain coherent, and so on).
What language users normally do know in advance, however, is the
overall topics or themes of the discourse (or discourse fragment) they are
about to produce. In the semantic module, therefore, these overall topics or
semantic macrostructures play a fundamental strategic role — they allow not
only global planning (and global understanding) of discourse, but also the
management of a large amount of information over a longer period of
238
Discourse
speaking or writing (or reading). 3 Topics thus also allow language users to
make their discourses coherent and to announce to recipients what they are
going to speak about (which may be essential to get the floor or attention in
the first place). It is also for this reason why many discourse types typically
express their 'upcoming' topics up front, for instance in various forms of
announcements, summaries or headlines. In comprehension this will allow
the recipients to activate or build up the top structure of relevant mental
models. They will know what the discourse will be 'about', and this
knowledge facilitates further understanding.
Under the overall control of topics, the semantic production module
finally produces the actual 'meanings' of discourse, in the form of a locally
coherent sequence of propositions. This happens by selecting the more
detailed, lower level propositions of the model a speaker has about an event.
As suggested, the context model specifies which lower-level information
will be relevant for actual expression, and which information may be left
implicit. Besides the construction of minimal local coherence, the speaker
may also shape its semantic representations following a number of strategies
that allow the differentiation of importance, focus, foregrounding and other
forms of information distribution and emphasis. Obviously, this linear
production of the meaning(s) of a discourse is also a strategic, ongoing
process, in which constraints from other modules may influence current
meaning production: ongoing thought and inferences, current perceptions
and experiences, interpretations of reactions of recipients (in oral discourse),
as well as any change in the ongoing context model.
The formulation module
The formulation module takes the output of the pragmatic and semantic
modules and produces actual utterances in a given natural language, using
the various discourse rules, grammar and lexicon of that language. This
production process is exceedingly complex. It takes place in working
memory and also has a strategic nature, with continuous feedback from the
pragmatic and semantic modules. Production is linear, and proceeds word by
word, phrase by phrase, clause by clause, gradually translating units of
semantic representations, such as concepts or propositions, into lexical
expressions, in their appropriate grammatical order. Although mistakes can
be corrected, the strategic nature of discourse production allows for a lot of
'imperfection', as long as the language user is being understood and speaks
or writes appropriately in the present context.
Specific semantic structures of the meanings to be expressed may thus be
mapped onto specific syntactic structures (word order, clause structure);
agency may for instance be embodied in the expression of a lexical item in
first ('topicaP) position and as the subject of the sentence; relations between
propositions may be marked by conditional or functional connectives, and
main topics may be placed up front in headlines.
From cognition to discourse
239
From these examples we also see that the formulation module not only
calls upon the grammar and the lexicon, but also on other discourse structure
rules and strategies, such as the structures of stories or news reports. To
write a news report, a journalist knows that the report should have a headline
or will expect someone else to write one for the report) and a lead as initial
discourse categories, and that these should express the most relevant
information in the present context, namely, the topics being constructed for
the present discourse.
Finally, when combined with lexical expressions, semantic 'content'
íerived from the semantic module (and its event model) and controlled by
the pragmatic module (and its context model) will be actually expressed in
talk or writing, following the usual phonological rules, for example of
intonation, or the graphical rules for the current genre.
Producing ideology
The details of these respective modules are not relevant here. 4 My brief
summary is merely intended to give an idea of how mental representations
'get finto' actual text and talk. Conversely, they also explain how the
understanding of text and talk may contribute to the construction of mental
representations. The question now is how ideologies may interfere in these
processes. Again, there are several ways in which this may happen.
Direct expression
Since under special conditions all accessible mental representations are
available for direct expression, ideological propositions may sometimes be
expressed directly. That is, if the contextual constraints of the pragmatic
module allow this, the semantic module may directly select the relevant
ideological propositions as input for the semantic representations (meanings)
of discourse. This is for instance the case in explicitly ideological discourses,
such as propaganda, theoretical analysis, and for discourses in which
ideological explanation, justification or legitimation is at stake. People in
that case primarily speak as group members, and express what believe
in. In a dispute with the unions or the government, for instance, managers
may directly state that 'the market does not want any government interferencé . Obviously such direct expressions may be combined with more
particular ones, such as personal experiences. Moral conclusions of stories
about minorities, for instance, may express the negative group evaluation
that 'we are not used to that heré . Given the abstract and general nature of
ideological beliefs, also the meanings (and their formulations) need to be
general and abstract, and feature generic concepts and expressions.
Instantiated direct expression
Ideological beliefs may also be expressed through instantiation (or specification) in mental models in episodic (personal) memory of the general
240
Discourse
propositions in social memory. For instance, instead of talking about
markets and governments in general, thus, specific managers may express
that they do not like the interference of this government. Disclaimers often
feature such instantiated direct expressions. As a strategy of positive selfpresentation, people may begin negative statements about minorities, by
saying'I have nothing against minorities, but ...'. The first clause of this
type of utterances, realizing an apparent denial, instantiates for the current
speaker the general opinion, derived from a non-racist ideology, that one
should not say negative things about minorities. That is, as soon as general
moral mies, attitudes and ideologies are applied to the present context and its
participants, we have an example of an instantiated direct expression of
ideologies. In formal terms, this means that variables are replaced by the
constants (for participants, time, place, etc.) of the present context.
Direct expressions of ideological attitudes
What has been said for the direct expression of ideologies also applies to the
direct expression of the domain-specific attitudes controlled by ideologies.
For instance, under the control of a racist ideology, group members may say
that they are in favour of a restriction of immigration of non-European
people. As with the expression of ideologies they may do so in general,
abstract terms, and use the group-reflexive or they may do this in the
instantiated form and use personal pronouns referring to specific
participants or subgroups.
In all the cases mentioned aboye, the information of social representations
is directly combined with the constraints of the pragmatic module and
entered into the semantic module of the discourse production unit. Conversely, in interpretation and (critical) analysis, discourse produced in this
way may be understood as explicitly expressing or indirectly signalling such
ideological beliefs. We should not forget, however, the possible constraints
of the context model. Both recipients and analysts should know that such
expressions may be made for a number of special social reasons, such as
social compliance, or the realization of specific goals (e.g. get a job). That is,
the pragmatic module may require people to be polite, tactical, or otherwise
forced to hide their'real opinions'.
Event model expression
Most discourse is about concrete experiences and events, and therefore
derives its information from event models, as described aboye. Ideological
and attitudinal group beliefs in this case may be instantiated and applied to
concrete personal situations. Instead of general opinions about noninterference of the government in the market, we may for instance have a
concrete news story in which specific managers resent a government policy
to have them register the number of members of ethnic minorities groups in
firms in order to get information about minority employment and discrimina-
:-
From cognition to discourse
241
tion. Similarly, beliefs about the criminal activities of a Turkish neighbour,
which may or not be based on personal experiences of the storyteller, may
similarly be an instantiation and application of the general ideological
opinion that minorities are criminal. Once part of the event model (the
personal construction or interpretation of the event), this personal opinion
may be used as input to the semantic module. Under the constraints of the
context model, people may or may not include such opinions in the semantic
representation of a story or an argument.
Context model expression
Since all models may thus be ideologically influenced, this is also the case
for the context model. People may represent co-participants in a negative
way only because they are members of specific social groups. Their
intentions, goals and actions may enact beliefs derived from ideologies and
attitudes, for instance when they directly intend to derogate co-participants.
Thus, intentionally or unintentionally issuing a command instead of a polite
request in a context where Chis would not be appropriate, may count as an act
of discrimination. The same is true for the contextual constraints on
deference and politeness, and other interactional conditions of appropriateness.
Negative representations of other participants in many ways influence the
semantic and formulation modules. Beliefs of events models that normally
would not be expressed because of contextual constraints of politeness or
non-discrimination may now be admitted to the semantic representation of
the discourse. Similarly, also various expression structures may directly be
affected by such 'biased' context models, for instance in the use of impolite
pronouns or intonation, and lexical items may be selected that signal
negative opinions about people spoken to or spoken about.
The fundamental role of context models in shaping (and interpreting)
discourse by the participants of communicative events, should again wam us
that a 'dame ideological analysis of discourse is theoretically and practically impossible. We should always know the details of the context in
order to know whether and what type of ideological control is at work.
Indeed, the 'same' statement in one context may have an ideological source,
which it may not have in another context — depending on the speaker, group
membership, intentions and goals, circumstances and so on. People may for
many reasons want to conceal their personal or group beliefs, or they may
express beliefs they do not have. They may feign, fie, dissimulate, be ironic
or metaphorical, and in many other ways say what they do not mean
fiterally. Thus, contexts in many ways 'key' the meanings and expressions of
discourse, and, without knowledge of that key, we are unable to understand,
infer or criticize their discourse or communicative act. In the studies of
specific ideological and discursive strategies in the following chapters, this
important warning should be heeded.
242
Discourse
Concluding remark
The discourse production processes briefly discussed in this chapter appear
to overlap with the ways underlying ideologies control other social representations, such as attitudes, which in turn may influence the opinions of
context and event models, which finally define the contents for the modules
of discourse production. Discursive and ideological production and reproduction thus run parallel, but at the same time it has been shown that the
expression of ideologies usually requires several stages. Few discourses are
wholly ideological in the sense that they express 'puyé group ideology.
However, general ideological opinions may of course be 'applied' in specific
models and thus provide the ideological basis for actual discourse production.
25
Persuasion
Influencing the mind
4lthough the theoretical frarnework proposed in the previous chapters
explains both the expression and the reception side of the relations between
discourse and ideology, the theory has focused mostly on the ways ideologies are expressed in text and taik In this and the next chapter, I shall take
the other perspective and examine some of the discursive and cognitive
strategies of the ways ideological discourse may be persuasively used in the
formation or change of ideologies. Thus, assuming that members of a group
effectively express their ideologies in their discourses, we now need to know
the 'effects' of such discourses on the minds of both ingroup and outgroup
members.
As is the case for the other fundamental notions studied in the previous
chapters, the notions of 'effects' and 'persuasion' have generated an
enormous literature in social psychology and mass communication research.
The empirical results of all this work, especially in the traditional effectsresearch in mass communication studies, have at best been rather inconclusive. The mass media, which undoubtedly also are the main means of
ideological reproduction in contemporary societies, have variously been
described as powerful or as rather powerless in influencing the minds of the
audience. Some research emphasizes that at most they are able to set the
agenda of public discourse and opinion. That is, they may not tell people
what to think, but they are quite effective in influencing what people will
think about. 2
This is not the place to review this vast research tradition. One major
problem of much earlier research is its theoretical inadequacy in the account
of the two main domains involved in the notions of 'effect' and 'persuasion',
namely, discourse and the mind. That is, in order to be able to say something
analytically acceptable about the influence of discourse, one needs an
explicit theory of the various structures of text and taik and their contexts, as
well as a cognitive theory of discourse comprehension and other mental
representations involved in understanding and cognitive effects. It is only
recently that the latter problem has begun to be dealt with, whereas discourse analytical approaches still remain scarce in effects and persuasion
research, mainly as a consequence of the unfortunate insularity of empirical
(read experimental) research in much social psychology, especially in the
USA.
244
Discourse
Another problem that is relevant for my discussion is the confusion about
the kind of mental representations involved in the processes of change
involved in the persuasive effects of discourse. No distinction has usually
been made between opinions and attitudes, nor between personal or contextual changes and long-term socially shared ones, as is the case for
ideological influence. Although there is work on attitude change in more
natural settings, most experimental work focuses on short-term, experimentally controlled changes observed in the laboratory. Moreover, much of the
work is also marred by the strange division of labour between cognitive and
social psychologists, of which the first deal with knowledge and learning,
and the second with opinions and attitudes, although in both cases the
processes and representations involved are closely related.
Different types of influence and persuasion
Against this background, then, the analysis of persuasion must be based on
the theory of cognition and text processing summarized in the previous
chapters. The implications of this approach for the study of the ideological
influence of discourse are the following.
1 Discourse understanding and influence is a complex process that is a
function of both the structures of discourse as well as of the mental
processing and representation of recipients. That is, whether or not, and
how, people are influenced by talk and text also depends on what they
already know and believe.
2 Discourse understanding not only involves the processing of structures
of text and talk, but also, and very crucially, those of the context as it is
subjectively construed by the recipients in their context models. In my
terms this means that the construction or change of any mental representations of events is a function of the contents and structures of
ongoing context models. One well-known notion to be explained in such
terms is for instance that of 'credibility'.
3 Although the relations between factual beliefs (knowledge) and evaluative beliefs (opinions, attitudes) are quite complex, it may be generally
assumed that discourses have an influence on evaluative beliefs only
when they are at least marginally understood. In other words, persuasion
presupposes comprehension. On1y in very specific circumstances may
people be persuaded by discourses they do not understand, and, even
then, at least partial understanding is a minimal condition of opinion
formation and change.
4 If persuasion is defined as a process in which people change their
opinions as a consequence of discourse, it is crucial to make a distinction
between different kinds of evaluative belief and hence between different
kinds of persuasion. Thus, a distinction has been made between personal
opinions and socially shared opinions. The first are represented in both
event models and context models, stored in episodic (personal) memory;
Persuasion
245
the latter in social representations, such as attitudes and ideologies,
stored in social memory. Moreover, one should distinguish between
particular and general or abstract opinions. Most socially shared attitudes
are by definition context-free, and hence abstract and general. Personal
opinions may be both particular and general: I may dislike my boss
today, I may dislike my boss in general, and I may dislike all bosses.
Socially shared opinions may also change, but because they are acquired
fairly slowly, also such changes take time. In sum, unlike much
traditional work on attitudes and attitude change, I do not simply
collapse all evaluative beliefs into one undifferentiated category of
'attitudes'. The concept of attitude is used here in its original sense of a
(set of) socially shared opinions.
5 From these distinctions it follows that discourses may variously affect
these different types of evaluative beliefs. My taik now may temporarily
lead to the formation or change of a particular opinion of my recipient
today, or may have more general effects: opinion change may be more
permanent, or it may affect more general and abstract opinions of the
recipient. And finally, a large number of discourses may have persuasive
effects on a large number of group members and thus gradually construct
or change their social representations, as is typically the case for the
more 'structural' learning from educational discourse or media discourse. Obviously, the acquisition or change of ideologies belongs to the
latter type of discursively based changes of the 'social mind' shared by
the members of a group, society or culture.
6 In principie all discourse may have ideological effects, whether or not it
expresses ideologies explicitly or implicitly. In practice, however, we
often limit research into ideological influence to those discourses that
express ideologies. That is, if we are interested in the production and
reproduction of ideologies, we usually will focus on the presence or lack
of ideological effects of ideological discourse.
7 It also follows from the theoretical distinctions made aboye that the
analysis of all processes of effect, influence or persuasion needs to relate
detailed structures of text and context, with those of short-term discourse
processing, as well as the details of mental representations both in
episodic (personal) and in social memory.
Ideological influence
After this brief summary of some of the main principies presupposed in the
analysis of the ideological influence of discourse, let us now tum more
specifically to the discursive and cognitive structures and strategies involved
in the formation and change of ideologies as a result of verbal communication.
Besides verbal discourse also other semiotic messages (images, photos,
movies, etc.) as well as other social practices may have ideological 'effects'
246
Discourse
on social members. Indeed, many sexist practices as well as ideologies of
men may be inspired by observation, interaction and watching movies, and
not just by male ingroup talk and text about women. Yet, in the test of this
book, I shall take such other semiotic and 'practicar influences for granted
and focus on discourse, with the understanding that the basic processes of
ideological influence involved are very similar.
Cognitive conditions
Discursiva influence on ideologies presupposes a number of cognitive
conditions. Before ideologies are being acquired and changed, people
already have vast amounts of factual and evaluative beliefs, represented in
the ways explained before. During socialization, education and peer group
interaction, thus, personal knowledge of members of groups and cultores
about concrete people, events and facts is thus gradually extended with
socially shared beliefs. That is, people learn that other people in similar
circumstances have the same or similar beliefs, or, vice versa, they learn to
accept (or reject) what they are told by others. In other words, the acquisition
of new ideologies by competent language users and social members does not
take place 'on a clean slaté .
Thus, we may generally asstime that ideological persuasion is facilitated
by lacking social and political knowledge, if recipients have no alternative
opinions, and if ideological propositions do not obviously clash with their
personal experiences?
More specifically, social members have gradually learned to distinguish
between (true or falsa) factual beliefs and evaluative beliefs, that is, between
beliefs that in principie should follow or be made plausible by truth criteria,
and those beliefs that represent people' s personal evaluation of situations,
events, objects, people or their properties in terms of cultural shared values.
As suggested, they also have acquired the cognitive competence that allows
them to distinguish between their personal opinions and those of others, and
that groups of people sometimes have the same or similar opinions.
And finally, people have leamed that their own knowledge as well as their
beliefs, and those of others, may change as a consequence of what others tell
them. For the change of knowledge this usually means that facts must be
supported by commonsense (or scientific) truth criteria, such as those of
reliable observation, correct inference or communication from credible
sources. For opinions on the other hand, change is usually related to 'good
arguments', based both on facts as well as on basic values about what is
good or bad, right or wrong. 4
In sum, the acquisition of ideologies takes place in a rich and welldeveloped social and cognitive environment: people know that others may
have the same or different opinions about the world and that such opinions
may be influenced by discourse. In a later stage they learn to discern that the
distribution of opinions of 'others' is not random, but that various >
people also tend to have various 'kinds' of opinions, and that many opinions
Persuasion
247
hang together. As is the case for all social leaming, they may acquire such
insights either indirectly on the basis of their own observation and inter:iction, as well as more directly through discourse: they may hear from their
parents, friends, children stories or TV that people not only have opinions
that may change or remain more or less the same, but also that group
inembership may be related to what people typically do or ought to think.
[ndeed, the early acquisition of gender knowledge and roles is an example
where boys and girls learn that they may have different opinions precisely
because they are boys and girls.
At the same time, they leam to understand that many of the earlier general
opinions (e.g. about boys and girls, children and grown-ups, or in general
'ú and 'them') seem to be relevant for the evaluation of many different
situations and events. It is at this point that the more complex attitudes they
have acquired during adolescence begin to crystallize into fragmentary
ideological systems with which they can personally identify.
Social conditions
The acquisition of social representations not only has a number of cognitive
conditions as informally summarized aboye, but also social conditions.
People have learned that social interaction in general and discourse in
particular is relevant in the way they and others acquire or change their
opinions. They know they have to defend theirs against others, and they
understand that others give arguments for their own opinions. And they have
understood that opinions are not only personal but may also be related to
group or category membership (being a boy, being a girl, or being a child).
[n sum, they know that opinions are often about social events or issues, often
y hared or disputed by others, acquired or challenged in social interaction,
and tied to social groups and different for 'o' and 'them'.
Obviously, such knowledge about opinion acquisition and change has its
empiricaP social base in the many forms of social interaction, communicadon, and group relations of which social members are part. That is,
ideologies just like other social representations are both a cognitive as well
as a social construction — they are not only mentally shared with others as
forros of social cognition, but also socially produced with others as group
members.
Ah this also applies to the cognitive and social dimensions of discourse
md its ideological influences. Members have acquired the social competence
not only to understand opinions of others, but also that these are typically
expressed in text or talk, and often in the forms of arguments. They know
that people may persuasively express both their own opinions as well as
those of the group or organization they belong to. Everyday discursive
interaction in which such opinions are being expressed are themselves often
part of broader social arrangements, organizations or institutions.
Children read or hear stories and watch TV and know that opinions may
be expressed by politicians speaking or quoted on TV or in the newspaper;
248
Discourse
and the same is true for the expressions of opinions by priests in sermons in
church, by teachers in lessons at school, or by fathers or mothers at honre.
Many of these opinions seem to recur in the same social situations,
expressed by members of the same group (other politicians, other teachers,
other parents, other girls) and it is in such social contexts, then, that groups
and systems of social beliefs tend to be associated to social structure, to
social groups, to social interests, and to self-serving talk and text of
members of such groups. In sum, through quite complicated processes of
social perception, interaction, communication and discourse, group members
gradually acquire the very notion of group attitudes.
The social and cognitive complexity involved here suggests that the very
notion of ideology, which is hardly a commonsense notion for younger
children, as well as the ideologies thernselves, are only gradually acquired
during adolescence. Indeed, the definition of ideology in terms of a complex
schema of categories defining the evaluations of the own group and its
properties (identity, activity, goals, norms, group relations and resources)
suggests that people will only acquire ideologies when they have leamed
what it is to be a group member. That is, from thinking in tercos of '1', they
have to learn to think in tercos of 'wé and 'them', distinguish a number of
group differences, identify with the group, participate in its activities, share
some of its goals, be subjected to its norms, values and tules, have
participated in inter-group interaction and conflict, and have been given (or
denied) access to social resources.
As such, these ideological schemata and the social conditions of their
(social) acquisition need not be acquired only in later adolescence: many of
the social experiences, social groups and social relations involved, children
already Nave acquired for age and gender — they know that grown ups often
have different opinions than children, and that boys/men and girls/women
also may have such differences, and children will thus have learned to
identify thernselves as children, have acted as children, defended their
interests or special resources, and so on. In other words, even when sociopolitical ideologies are acquired much later, the social conditions of primary
and secondary socialization are such that children already learn at an early
stage the relevant cognitive and social conditions of group membership and
the ways in which such membership is related to opinions. 5
Opinion discourse understanding
Under the social and cognitive conditions summarized aboye, social members are routinely and daily confronted with many types of discourses that
express socially relevant opinions. Por such discourses to have implications
for the formation of ideologies, they need to be understood in general, and
need to be understood as expressing opinions in particular. This does not
mean that discourse expressing factual beliefs does not play a role in
ideology formation. It does. We may daily learn about the hand facts of
killings in Bosnia, and may ourselves associate with these facts the evalu-
Persuasion
249
ations that may sustain or challenge for instance actitudes about Serbs,
Croats and Muslims, or about armies, or ideologies about ethnic conflicts or
about pacifism in general. Although obviously the media accounts of the
events in Bosnia are replete with critical opinions, which also will influence
our own attitudes, this example also shows that the formation of specific
ideologies need not take place through explicit opinion discourse. It may be
sufficient that people get what they see as the 'facts', and give their own
.[personal or socially shared evaluation of them on the basis of specific values,
in this case those of non-violence, or those of defending the weak against the
strong.
With this important caveat in mind, however, the opinions that are being
inferred from discourse are often pre-formulated in those discourses themselves. Understanding such opinion discourse has two different cognitive
consequences. People represent the events (such as those of the war in
Bosnia) in their event models, and at the same time represent the opinions
,about these events, also in the event model. On the other hand, they may
represent the events in the event model, but the opinions expressed by the
text as those of the speaker or writer, and then store those opinions in the
speaker/writer model that is part of che context model.
In che second case, che recipient may or may not agree with such opinions,
hitt merely represent them as the opinions of a particular writer or speaker.
That recipients construe such models of speakers/writers may be concluded
from the fact that recipients are usually able to reproduce later the opinions
of the speaker/writer. When generalized, these models may later even allow
recipients to conclude that the speaker/writer is a pacifist or a militarist, pro
or contra the Serbs, and so on, even without remembering the concrete facts
of the event model.
. If opinions are stored with the mental representation of che events
themselves, that is, as part of the event model, we might assume that the
opinion is provisionally accepted or adopted by the recipient. In the same
way as the discourse being understood may be evaluated as more or less
factual and as probably true or false on the basis of truth criteria, arguments
and what recipients already know or believe to be che case, also the opinions
expressed in discourse may thus be evaluated. If they match the general
personal or social opinions of the recipient, then the opinion may be
1 provisionally adopted and associated with the event in the event model. If
che result of this evaluation process is negative, then che opinion may simply
be attributed to the speaker/writer and stored in the context model, as
explained aboye. Probably the same is true for the factual beliefs expressed
in the discourse as well. That is, if the recipient does not believe that what is
said is true, then it seems pointless to construct a 'model of the events',
because the events don' t exist in the first place. Instead, the speaker/writer
will also be attributed with what he or she 'claimed' to be the case, and no
event model is then constructed.
Although this solution for the well-known problem of 'acceptancé of
bpinions seems elegant, it also has its drawbacks. Mental models were
?_
250
Discourse
introduced in cognitive psychology in order to account for a variety of
problems in (discourse) understanding. That is, beyond semantic representations, understanding a discourse involves the construction of a model. When
people are able to construct at least a fragmentary model of what the
discourse is about, we say that they have (at least pardy) understood a text.
The question of truth or falsehood is not a condition of such understanding.
Indeed, since we were children we have learned to understand and construe
models for myths, fairy tales, lies and fiction. That is, a model represents any
kind of event, fictional or real. In fact, one of the reasons to introduce the
notion of model in the first place was that such models are constructed by
people whether or not they believe or know that the events being talked
about are true or false.
If this is the case, we should conclude that at least for the representation of
'factual' (true, false, fictional or not) beliefs, language users construe
models. That is, even 'false' discourse needs to be understood, and the way
to do that is to build a model for it. It will not do to simply construe 'what
was raid', that is, a semantic representation, and associate that with the
'discourse model' that is part of the context or speaker models of the current
situation.
Now, if this is true, we still need to account for an independent assessment
of the truth or falsity of the events represented by the model. Again, the
easiest way would be to simply 'tag' the model as being truthful or false (or
assign it a probability value), as a result of the evaluation procedure that
compares the 'facts' of the model with other (true) models or with
instantiations of general, shared 'certified' knowledge. This would also mean
that if such a tag were no longer accessible later, people might erroneously
'believe' what they once represented in the model, a condition that is quite
familiar in media reception studies. Another option would be to store
(believed) models in a separate memory location, and mark that location as
(personal) 'knowledge'. The advantage of such a solution would be that such
a separate knowledge reservoir would be more easily related to socially
shared, 'accepted' knowledge.
Theoretically, however, these two ways of representing subjective truth
and falsity would be practically 'notational variants', as linguists would say,
although empirically one or the other proposal might have different processing consequences. The main point is, though, that all models get stored in
episodic memory, and most of them will get evaluated during processing (or
sometimes later), and then marked as being (more or less) truthful.
The same argument, however, does not apply to the representation of
opinions. Indeed, opinions are not properties of facts but of people, so
opinions are not stored 'with' the event, unless they are (like truths) the
result of the evaluation procedure of the recipients themselves. That is, if the
recipient represents a fact like the rapes of women in Bosnia, then it will be
her or his opinion being associated in his or her own model of these facts
that gets represented. Opinions of speakers or hearers on the other hand do
not thus attach to the events and the event models, but to the speakers/
Persuasion
251
writers and their models as part of the context model. This seems to be
confirmed by the fact that for instance readers of news are able to construct
a model (their model) about an event independently of the possibly biased
opinions of the speaker/writer. Indeed, they may even disregard the biased
style of the discourse, and reconstruct the model contrary to the persuasive
intentions of the writer/speaker. I shall return to this notion of 'preferred'
interpretations later.
Unfortunately, there is as yet no theoretically satisfactory and sophisticated way to represent what we are dealing with in the first place: opinions.
It was decided to simplify matters for the moment and represent these as
'evaluative' propositions, that is as propositions with an evaluative predicate, where such a predicate is any concept that is derived from some social
or cultural value. But we have seen in Chapter 11 that the difference
between'factual' and 'evaluative' propositions and predicates is more
complicated. Although many predicates are generally treated in a specific
ociety and culture as being evaluative ('beautiful',
'right', 'wrong',
etc.) or as factual (e.g.
, 'stoné , 'pape? or 'car'), there are many
others where it depends on the perspective, values and indeed the ideologies
of the group members whether these are factual or evaluative notions (e.g.
'thief,, lerrorise, 'heavy' or 'pollutioñ).
Given this theoretical uncertainty about formats of representation, we
have at the moment no other alternative than to represent opinions in models
as evaluative belief propositions. But it should be added that this implies
precisely what ideologies are supposed to do, namely, that some people will
represent as models of'facts' what others represent as context models of
opinions of other people. This nicely ties in with the proposal that the kind
of 'bias' of mental representations as a function of different ideologies is
exactly what it is, namely, a differently organized system of models. We
may assume that various processing tasks, including the use of models in
discourse comprehension and production will be affected by such different
representations. That is, in the various structures of discourse, as well as in
processing such discourse, it should become clear whether an event is being
represented as truthful, or whether it is represented as false, and especially it
should show whether my opinions about such an event are par( of my model
of the event, or whether they are attributed to the speaker or writer. In the
latter case the opinion is represented in the model I have about the models of
the speaker or writer, and that representation is part of my context model and
not of my event model.
What was just proposed also shows how opinions about contexts are
being processed. That is, people do not only construct models of speakers or
writers (with their opinions about them) on the basis of what these say, but
also on other grounds, which have been discussed before, such as group
membership, appearance, non-verbal activities, and so on. The same is true
for the rest of the context, such as the ongoing communicative event as a
whole, the setting, props or circumstances, and so on. Obviously, these may
play a crucial role in the construction of event models and opinions. The
252
Discourse
familiar insights of credibility research obtain here. Thus, when a statement
about an event is made by someone who is known to be a specialist in the
study of such events, then such statements will be more credible than those
made by a non-specialist, unless other information (like special personal
interests of the specialist) overrides this truth criterion.
In other words, contexts, or rather subjective interpretations of context,
that is, context models, provide the resources being used in the application
of epistemic evaluation of discourses in the construction of event models.
People draw on their personal and social knowledge as well as on what they
know about the context (identity of the speaker, etc.) to decide whether what
is being said is more or less truthful. We have also seen that this also
explains the ideologically biased evaluation of the context (and hence of the
discourse) — if for racist reasons blacks are deemed to be less competent or
truthful, whites may assign a lower truth value to what blacks say. That is,
perceived group membership influences the construction of both context
models and event models, including the opinions and overall (truth) evaluations being assigned to them.
Generalization and abstraction
Having construed the event and context models featuring the opinions
derived from opinion discourse or construed by recipients as their own
opinion about the events or the context, other strategies will be applied to
make such opinions more useful for social members. That is, opinions
should also be relevant in other situations, and in the judgement of other
events and contexts. This requires decontextualization, abstraction and generalization, as described before — models of particular events and contexts
will be abstracted from in such a way that they may be used in the •
understanding and evaluation of other events. This may yield general
personal models, representing the personal experiences and opinions of each
person, but also social representations that are shared by others. For my
purpose, especially this latter strategy is relevant.
Again, little is known about the details of these strategies and under what
conditions they take place. For social group members to know that specific
factual or evaluative beliefs are shared by most or many other members, a
process of 'normalization' should take place — own beliefs, based on
personal experiences, need to be compared with those of others. This will
again usually require discourse — speakers belonging to a group who are
talking to other members (or reading texts from other members, e.g. in the
press) construct models of their interlocutors and their beliefs, and may
generalize such, models to social representations featuring the shared beliefs
of the own group. A variety of contexts, speakers, and circumstances as
properties of contexts, as well as specific features of discourse, such as
presuppositions, may thus suggest to group members that it is apparently
'generally accepted' that such or such is true or false. We see that the
abstraction and generalization of context models, that is, decontextualiza
Persuasion
253
mon, precisely provides the crucial criterion for the transformation of
personal knowledge into social knowledge.
The same is obviously true for opinions. If social members repeatedly
represent many other group members as expressing a specific opinion, they
may generalize and assume that this is a typical opinion of the group as a
whole. This holds true both for the generalization of ingroup opinions as
well as for outgroup opinions, although ingroup opinions may be encountered more often, be found more credible, and so on, and therefore more
easily acceptable than those of outgroup members. Inter-group perception
and differentiation thus also takes places at the level of opinion differentiation. Our facts or opinions may be sufficient reason to reject, a priori,
those of others, disregarding an 'independene evaluation of their validity. In
fact, group differentiation may be based only on the perception of different
;social opinions and not on other social membership criteria.
Finally, social opinion clusters (attitudes) may be further generalized and
abstracted from as ideologies, as described earlier. In this case, further
'decontextualizatioñ regards specific social domains or circumstances. Por
instance, women may acquire a number of relevant attitudes, for example
>; 'about equal pay, glass ceilings, child cace or abortion, and then abstract from
the various roles (and inequality or lack of autonomy) in situations at work,
the family or in politics, and derive the general ideological propositions that
represent what these different social situations have in common. Theoretical
and empirical details of these processes are , as yet unknown.
It was suggested earlier that the acquisition of ideologies need not be
indirect and based on models, but may also be direct, that is, based on
general statements about social representations and ideologies in discourse.
Instead of personal experiences and opinions, thus, social members may be
confronted with explicit attitudinal or ideological discourse and derive
relevant opinion propositions directly from this discourse, without the
intervention of models. Since no 'facts' sustain such social representations,
contextual conditions are crucial — speakers/writers need to be very credible
before people accept their general statements as valid. Again, decontextualdation may operate here — the same statements are being made by many
other ingroup members, so that such consensus information alone will
cnhance credibility. Yet, social members may still want to evaluate such
general statements with respect to their other social representations, and may
then accept them as valid when they are consistent with these other
tepresentations, suspend judgement when there is no consistency, and reject
them as biased when they are inconsistent with (many) other representations,
or eventually re-evaluate their current social representations. It is only this
latter process that one should call 'attitude changé .
We now have an approximate idea about the ways (opinion) discourse
influences the mind, which representations are involved, and how social
beliefs, including ideologies, may be confirmed or changed by discourse. We
have found that context models play a crucial role in the construction of
personal and social opinions, and the same is true for the event models and
254
Discourse
social representations of social members. Both sources are used as the basis
for the evaluation of discourse as valid or invalid. Ideologies may be
acquired 'empirically' but indirectly through the decontextualization of
particular and personal models to more general and more abstract representations, or they may more directly be formed by explicit expressions of
social beliefs. Evaluation of social beliefs usually requires decontextualization, however, so that even for explicit ideological discourse, repetition may
be needed by various and credible sources before an ideology is being
accepted. Most crucially, however, for all social representations, and especially their ideological underpinnings, is that they should'work'. That is,
they should be applicable in peoplé s everyday lives, in the adequate
accomplishment of social practices, in understanding such practices and
other people, and in the successful participation in discourse.
26
Legitirnation
What is legitimation?
Legitimation is one of the main social functions of ideologies. In classical
approaches, thus, dominant ideologies were usually described in terms of
their role in the legitimation of the ruling class, in particular, and the
dominant order, in general. In this chapter, I examine sorne properties of
legitimation and its relationships to ideology and discourse. 1 In philosophy,
law, and the social and political sciences, legitímation is a notion that has
been extensively studied. 2 However, in discourse analysis, it is much less
studied than, say, politeness or persuasion, although legitimation is a
prominent function of language use and discourse.
As may be expected in the framework of this book, legitimation will
primarily be defined in a discourse analytical framework. 3 It is obviously a
social (and political) act, and it is typically accomplished by text or talle
Often, it also has an interactive dimension, as a discursive response to a
challenge of oné s legitimacy. Pragmatically, legitimation is related to the
speech act of defending oneself, in that one of its appropriateness conditions
is often that the speaker is providing good reasons, grounds or acceptable
motivations for past or present action that has been or could be criticized by
others.
However, the communicative act of legitimation has several further
constraints, and does not, like defences, presuppose actual attacks or
challenges, but at most possible ones. Theoretically, legitimation is not an
illocutionary act at all, but (like argumentation and storytelling) a more
broadly defined communicative act that usually requires more than the
utterance of one single proposition. Legitimation may be a com4plex,
ongoing discursive practice involving a set of interrelated discourses.
Legitimating discourse is usually accomplished in institutional contexts.
Although people may perhaps be said to legitimate' their everyday actions
in informal conversations, such usage would probably count as being
derived from a more formal lexical register. In everyday informal talk, we
would rather speak of justifications, explanations or accounts. In all these
cases, the crucial point is that speakers explain why they did or do
something, and why such an action is reasonable or, in general, socially
acceptable. In such acts, we may expect arguments, that is, references to
reasons and to courses of action that had or have to be taken because of
contextual constraints, causes or opinions. Moreover, this family of commu-
256
Discourse
nicative acts is interactively engaged in especially, as we already saw, when
these reasons or these actions referred to are not obviously acceptable.
People justify or account for their actions mostly if they know or expect that
others might be puzzled or, more strongly, if others disagree, condenen,
challenge or attack them because of these actions.
Legitimation, then, is the institutional counterpart of such justifications.
That is, speakers are usually described as engaging in legitimation as
members of an institution, and especially as occupying a special role or
position. Legitimation in that case is a discourse that justifies 'official' action
in terms of the rights and duties, politically, socially or legally associated
with that role or position. Indeed, the act of legitimation entails that an
institutional actor believes or claims to respect official norms, and hence to
remain within the prevalent moral order.
Legitimation presupposes institutional restrictions of social power, as
defined by law, regulations, rights or duties that set the boundaries of
institutional decision-making and action. All those who have no absolute
power may routinely need to legitimate their action, although, for many (e.g.
face-keeping) reasons, even dictators will regularly engage in various forms
of legitimation. 6
Because of this institutional nature, legitimation may not be restricted to a
justification of official action, but even of the position, role or institution
itself. Accusations of illegitimacy often make normative inferences from
actions to the actor, or about his or her very incumbency in the position.
Indeed, in a democracy, a president of a country, when found to have
engaged, ex officio, in serious ilegal action, may expect to be impeached.
And the security services of a dictatorship accused of breaches of human
rights may be abolished by democratic governments because of their
illegitimacy.
These examples also suggest that legitimation is not only engaged in by
persons in some official position, but also by institutional actors, such as
organizations, official bodies, parliaments, and so on. That is, legitimation
may be a forro of collective action, and hence aims to justify the actions of
the institution itsel£
Legitimating discourses presuppose norms and values. They implicitly or
explicitly state that some course of action, decision or policy is luse within
the given legal or political system, or more broadly within the prevalent
moral order of society.
Given the relation between legitimation and institutional power, legitimation discourse is prototypically political. Those expected to legitimate
themselves are those who occupy or are appointed to public office and who
exercise power because of such office. In a state of law, Chis implies,
obviously, that they not only respect widely shared social conventions,
agreements and norms, but especially the law. 7
In the real world of politics, legitimation discourse may, however, be
expected especially when officials are accused of breaking the law, or when
they expect principled opposition against their decisions, policies or political
Legitimation
257
action. Indeed, legitimation may not be necessary in normal courses of
events, in routines, and when no challenges to institutional power or
authority are imminent. They become imperative, however, in moments of
crisis, when the legitimacy of the state, an institution or an office is at stake.
Legitimation, then, becomes part of the strategies of crisis management, in
which ingroups and their institutions need self-legitimation, and outgroups
must be delegitimated.
Note that the concept of legitimation used here has a top-down direction:
elites or institutions legitimate themselves especially 'downwards', that is,
with respect to clients, the citizens, or the population at large. There is also
a complementary form of legitimation, which is bottom-up, and involves the
legitimation of, for example, the state, elites or leaders by the 'masses'. For
instance, it has often intrigued social scientists why many forros of oppression and inequality are so often accepted or condoned, or even normatively
approved by people in subordinate positions. 8 One explanation for such
approval is 'equity': people often think that their subordinate position or the
dominance of the elites are deserved because of their respective actions or
performance. The criteria for this kind of self-evaluation, however, are often
established by the elites themselves, so that in fact this form of popular
legitimation is rigged from the start.
Legitimation and ideology
Within these succinctly summarized general principies of legitimation, we
now need to examine what the role is of ideology in such acts of
legitimation. Indeed, how can ideologies be an Instrumene of legitimation?
We have seen that legitimation presupposes moral or legal grounds for the
judgement of offlcial action, such as norms, values or formal laws. In our
analysis of ideology, we have seen that ideologies, as the basis of the social
representations of groups and their members, also presuppose norms and
values. For specific groups, thus, ideologies provide the foundation of
judgement and action, and hence also the basis for group-related legitimation. Thus, democratic ideologies provide the basis for judgements about the
legitimacy of 'democratic action
Similarly, xenophobic groups or parties may engage in racist actions, but
usualll deny that such actions are racist, and hence outside of the moral
order. Instead, they will claim that it is 'natural' to make a distinction or
even to establish a hierarchy between Us and Them, to accord priority to Us,
or to give preferential access to symbolic or material resources because of
Our blood, soil or innate characteristics. Racist ideology, self-servingly
appropriating general social norms and values about precedence and rights
of the ingroup, thus embodies the basic principles of the shared opinions that
control racist actions as well as their legitimation." As is the case for
justiflcations and accounts in general, socially shared representations, and
. 9
258
Discourse
especially the evaluative ones, provide the grounds for judgements about
what is right and what is wrong, good or bad.
In sum, ideologies form the basic principies of group-internal legitimation. They do so by specifying the ideological categories of membership
criteria, the activities, the goals, the social position, the resources (or power
base) as well as the norms and values for each group. These norms and
values not only regulate and organize the actions of group members, but also
may be used to justify (or indeed to challenge) the social position of the
group in relation to other groups.
It is at this point where ideology and legitimation interact most specifically, in the control of inter-group relations, such as those of power,
dominance and resistance. Indeed, as we have seen, the classical approach to
ideology was to define them in terms of their role in the legitimation of
dominance.
Since, however, ideologies are by definition group based, and hence
feature propositions that are in the interests of the own group, their
consequences for group action may conflict with those of others. Indeed,
membership criteria, actions, goals, values or access to resources of one
group may be inconsistent with those of other groups. This means that to
legitimate group action not only for group-internal purposes, but also for
inter-group purposes, a group needs to show that its basic principies are just,
and possibly that those of the other group are wrong. Or rather, it may claim
that its basic principies are general, if not universal, and hence apply to
everyone.
To legitimate actions in a social conflict and in a situation of inequality in
which one group is or may be challenged by another, usually involves the
claim that these actions are within the general moral order, and hence not
justified only by partisan, self-serving grounds. Group ideologies may thus
be declared to be 'common sense', or principies that should be followed by
all social members, also those of other groups. As we have seen, persuasion
and manipulation may thus be combined with legitimation as soon as one
group tries to impose its ideology on another group or is able to have it
adopted by more subtle means. 12
Delegitimation
At the same time, this obviously implies that opposing groups, as well as
their basic principies (ideologies), will be delegitimated. Ideological and
social conflict thus take the form of a struggle not only over ideas, or over
scarce social resources, but also over legitimacy. Domination in this case
will crucially involve those strategies that are geared towards the delegitimation of internal dissidence as well as outside competition or 'threat'. These
strategies may themselves follow the categories of the ideological schema,
and thus challenge the very existence or identity of the other group, for
example as follows for the case of delegitimating minority groups, refugees
or other immigrants."
Legitimation
259
1 Delegitimating membership: they do not belong here, in our group, in
our country in our city, in our neighbourhood, in our organization.
2 Delegitimating actions, including discourse: they have no right to engage
in what they do or say, for example work here, or accuse us of racism;
criminalization of actions (e.g. 'ilegal entry').
3 Delegitimating goals: they only come here to take advantage of our
welfare system.
4 Delegitimating norms and values: their values are not ours; They should
adapt to our culture; We are not used to that here.
5 Delegitimating social position: for example, they are not real refugees,
but merely economic ('fakeD ones.
6 Delegitimating access to social resources: they have no priority to get
jobs, housing, work, welfare, education, knowledge, etc.
For each social group that is seen to challenge the dominant group(s) or
the status quo, the main identifying categories defining the group may be
delegitimated. Thus, for goal-defined groups such as social movements, the
strategy will focus on the delegitimation of their goals, as is the case for the
womeñ s movement or the peace movement. If for instance the goal is to
end patriarchy or sexism, this goal may be delegitimated by denying that
gender inequality is a major problem in society." For ideological opponents,
the basic ideologies will be attacked as being inconsistent with the dominant
values. And in the neo-liberal mover to abolish welfare, such access to a
crucial resource will be delegitimated by reference to the need to push back
the role of the state, and to emphasize the need for people to take their own
initiative to find a job.
These examples also show that strategies of delegitimation generally
presuppose norms, values and ideologies that are claimed to be universal or
widely accepted in society. Dominant groups will in such a case not openly
refer to their own interests, but on the contrary engage in arguments that
claim that their actions or policies are for the common good or are good for
the dominated groups themselves. This is for instance typically the case in
the political delegitimation of immigration and hence of afi immigrants. It is
not surprising that the most pervasive adjective in official discourse about
immigrants is that they are 'filegar. By thus portraying immigrants as
people who break the law, the strategy at the same time implies that they are
criminals, and place themselves outside of the civil society, so that immigration restriction, expulsion and withholding social services to immigrants
become legitimate.
Thus, in Europe, the elites will not refer to their own privileges when
opposing immigration, but will focus on the consequences for poor (white)
people in the inner cities, or may emphasize that it would be better for
immigrants if they would help to build up their own country.
In the same way, the peace movement may be delegitimated br emphasizing its violente, and hence violating the value of non-violence. l Unions, or
strikers may be delegitimated by focusing on the dire consequences for the
260
Discourse
economy (the common good) when their demands are realized, if not on
their illegal actions, their violence, their 'communise ideology, or the threat
to freedom (of the market). 16
Legitimation, delegitimation and discourse
It has become clear aboye that legitimation is a complex social act that is
typically exercised by talk and text. Strategies of legitimation and delegitimation are similarly discursive, and involve the usual moves of positive
self-presentation and negative other-presentation we shall farther examine in
the next chapter.
But as is the case for all social action, discourse itself may also be
(de)legitimated. This is a crucial strategy, because discourse was found to
have a primary role in the formation and change of underlying attitudes and
ideologies, that is, in persuasion. If the public discourse of any social group
can be controlled or delegitimated, a dominant or competing group can
establish hegemony over the symbolic domain, namely, the control of the
meanings and minds of the recipients of such discourse. In war, civil war,
revolution or social conflict, one of the main targets of attack will be radio or
television stations, or the exercise of censorship. And where coercive force,
prohibition or other legal measures are impossible or ineffective, strategies
of delegitimating or otherwise marginalizing opponent discourse will be
resorted to.
Strategies geared towards the delegitimation of discourse take several
forms. First, they may focus on the context of production, on access and use
of discourse, for example by challenging the legitimacy of communication
participants (who has the right to speak, or to speak for others?), speaker
roles, setting, goals, knowledge, expertise and so on. Newspapers may thus
deny representatives of 'illegitimate' groups access to the newspaper,
boycott press conferences, ignore press releases, or represen t leaders or
speakers of movements as unreliable sources in newsgathering.' 7
One very effective form of ideological speaker control is when dominant
groups are able to influence the minds of the speakers themselves, through
the interiorization of dominant beliefs, attitudes or ideologies. There are
many examples, for instance in the domains of class, gender or where
dominated groups have been confronted so consistently with legitimate,
official discourse, that they may accept that they are indeed inferior, deviant
or otherwise illegitimate. We have seen in the previous chapter how subtle
processes of persuasion and manipulation are able to create preferred mental
models of events. These models may then be generalized to more fundamental, shared social self-representations of a group. These will in tum
control the everyday judgements and social practices of the members of the
dominated group, in such a way that they are consistent with the interests of
the dominant group. Of course, this is the standard example of how
dominant ideologies work in the formation of 'false consciousness', and we
Legitimation
261
have seen that in the real world, such ideological hegemony is seldom
complete, given the many forms of mental and social resistance by dominated groups. Obviously, these forms of counter-power and resistance are
themselve§ again in need of legitimation, which itself is based on a counterideology.
Yet, given the close relation between ideology and social identity, such
ideological brainwashing may also affect the very self-confidence of whole
groups. This has often been observed for women and blacks confronted with
pervasive derogating discourse by men and whites, respectively. It is only
through raising group self-consciousness and ideological de-programming
that the effects of this form of ideological hegemony may be countered.
Second, once access to public discourse cannot be prohibited or denied,
opponent discourse may itself be delegitimated by many moves. These may
include, for example, citing out of context, focusing on negative or threatening elements in discourse, emphasizing the violation of common values,
or by framing such discourse in a specific way, for example through negative
speaker description - Marxise , 'radical', 'fundarnentalisf,, etc.).
Thus, of the speeches of 'radical' Nation of Islam leader Farrakhan, the
media will typically focus on his anti-semitic remarks, as they also did when
African American leader Jessie Jackson spoke of New York as Ilymietowñ . In this case, it may be left to the readers to draw conclusions about
the reliability and the legitimacy of the speakers of the others. Another
framing strategy is to use authoritative, and hence legitimate' speakers, for
example police officers or the mayor, in order to correct possible accusations
by minority groups after a 'riot'. Indeed, as I have found in my work on
racism and the press, minority representatives seldom are allowed to speak
alone, and hence function as the only source about ethnic events. This is
especially the case in crucial accusations, for example of racism. Not only
will these be presented as fundamentally doubtful, and hence between
quotes, but also they will never go unchallenged by the (white) authorities.
Finally, the delegitimation of opponent or dissident discourse by dominant
(political, media, etc.) groups and organizations may focus on the possible
effects of such discourse, and hence on the recipients. Of course, this may be
done, indirectly, by presenting speakers and discourse themselves as illegitimate, for example while being unreliable, violent, radical or deviant. Event
models and context models of recipients are thus persuasively oriented
towards negative representations of the 'illegitimaté speakers or to a
rejection of what they say as being true. But, even reception itself may be
obstructed, for example by programming broadcasts at times when the
audience is small, publishing items on inside pagel or inconspicuous places,
by jamming the airwaves, imposing duties on distribution of radical media,
preventing the public from listening to speeches, and so on.
Also in democratic systems that celebrate free speech, there are many
ways to delegitimate dissident or opponent discourse in many overt or subtle
ways. This essentially happens by preventing or impairing access — to the
media of public discourse, to fair representation, and especially to the minds
262
Discourse
of the audience at large. At the same time, of course, own group discourses
will be favoured in the opposite direction, and will have optimal access to
context, text and reception.
For the discussion in this book, this analysis of the (de)legitimation of
discourse is important in understanding ideological conflict and reproduction. If dissident or opponent discourse is delegitimated, and hence the
'normal' processes of communication and persuasion are impaired, also the
construction of altemative ideologies is made more difficult. In social,
political and ideological conflict and crisis, it is vital that members of the
ingroup, or members of allied or neutral groups do not become 'infected' by
the ideological virus of the opponents. Once such an ideology is allowed to
spread, it will increasingly control the social representations, models and
hence discourse and other actions of the population at large. In thát case, not
only may the ideological struggle be lost, but the social and political struggle
as well, if the persuaded others act in accordance with their new ideology.
The strategies of the virulent anti-communist witch-hunt by Joe McCarthy in
the USA are a prominent example of the forms of ideological delegitimation
described here.
As we have seen, strategies of delegitimation presuppose power and imply
dominance, that is, power abuse. In the domain of discourse and communication, such power need not merely be political or socio-economic. It may
also be symbolic. That is, dominant discourse may be presented as legitimate because it has authority and prestige, and hence is associated with
truth. 19 Thus, politics and especially the media and science exercise ideological control because their discourses are legitimated by the control over truth
criteria, such as information, evidence and expertise. If no counter-evidence,
counter-expertise or altemative information is (made) available by their
opponents, thus, such elite discourse is self-legitimating because of its
exclusive access to such symbolic resources as authoritative knowledge and
opinion.
Moreover, powerful elites also control the institutions that organize such
special access to knowledge, truth and opinion, such as universities, laboratories, think tanks, intelligence agencies, secret services, bureaucracies and
so on. That is, their authority defined in terms of truth claims may be
effective not merely by preferential access to public discourse or media
control, but albo by the 'incontrovertible' (reliable, scientific, etc.) evidence
that will back up such claims Thus, the strategies of legitimation are most
effective when they are able to establish the very norms, values and
ideologies by which both dominant and dominated groups and their actions
are judged. In the next chapter, we will examine some of the discursive
properties that may be brought to bear in such ideological legitimation and
control.
21
Discourse Structures
On levels, structures and strategies
Typical for a discourse analytical approach to ideologies and their reproduction is that ideologies are not símply related to undifferentiated forms of text
or talk, but mapped on to different levels and dimensions of discourse, each
with its own structures or strategies. These various properties of discourse
are the result of theoretical analyses and therefore may vary widely in
different approaches.
Thus, conversation analysts exclusively focus on spontaneous, everyday
dialogues, linguists on the grammatical structures of discourse, whereas
pragmatics focuses on more speciflc properties of action and interaction,
such as speech acts, illocutionary force or politeness strategies. Whereas
earlier 'text linguistics' in practice tended to study mostly written texts, most
other contemporary approaches, especially in the social sciences, have a
preference for the analysis of spoken discourse, sometimes with the implicit
assumption that 'natural' language use is essentially oral and interactive.
Psychology on the other hand favours the study of (written) text comprehension, probably also because this is easier for experimentation in the
laboratory.
It needs little argument, however, that both spoken and written/printed
forros of discourse are the object of discourse studies, and that diere is no
more or less 'natural' priority here, at least not for all cultures that have
writing systems. Any approach that uniquely associates ideologies or social
representations with the interactive, face-to-face social construction of
'meanings' is therefore by definition incomplete: ideologies are also
expressed and reproduced by written text. Indeed, when it comes to the
mass-mediated reproduction of ideologies in contemporary society, face-toface interaction may even play a less prominent role than textual or onesided spoken/visual communication by newspapers and television.
From the sprawling cross-discipline of discourse studies that has emerged
from anthropology, sociology, linguistics, psychology and other disciplines
in the humanities and social sciences, we may hardly expect anything else
but a large variety of approaches, theories, methods and their underlying
philosophies. In order to give some background to the chapters that follow,
let us briefly summarize some of the main structures usually studied in
discourse analysis. At the same time, I give a brief indication of the ways
ideologies may impinge on such structures during their communicative
Discourse structures
201
manifestations. Note, though, that these indications are merely illustrations.
A proper discourse analysis of ideological expressions of course would
involve a much more detailed and systematic account of relevant structures
and strategies.
Graphics
Neglected in virtually all approaches of discourse studies, and obviously
irrelevant for the study of spoken dialogue, graphical structures of written or
printed text are literally a prominent, while actually visible, property of
discourse. Apart from some semiotic work on images or textual graphics,
theory-formation in this field is still scarce, and analyses hardly go beyond
impressionism. Yet, litde theory is necessary to understand that variations of
graphical prominence may constitute a crucial element in the expression of
ideologies. Whether a news report appears on the front page or on an inside
page of the newspaper, high on the page or at the bottom, left or right, or
whether it has a small or a banner headline, is long, short or broad, that is,
printed over several columns, with or without a photograph, tables, drawings, colour and so on, are all properties of the graphical representation of
just one genre that may have a serious impact on the readers' interpretation
of the relevance or newsworthiness of news events. Many advertisements are
inherently associated with images, colours and other graphical elements, and
sometimes lack verbal text altogether. The visual element of TV programmes is crucial, and also includes special discourse graphics. Modem
textbooks have a graphical layout that is assumed to raise and keep the
interest of children and adolescente. And so on for a large variety of other
written or printed genres.'
Graphical structures may have several cognitive, social and ideological
functions. Cognitively, they control attention and interest during comprehension, and indicate what information is important or interesting, or should be
focused on for other reasons, and may therefore be better understood and
memorized. They may signal communication forms and genres, such as the
difference between a news report and an editorial in the press, or between
theory and assignments in a textbook. Socially, graphical structures, including photographs, have a large domain of associations, for instance with
groups, organizations and subcultural styles, as the difference between a
popular tabloid and a serious mainstream broadsheet shows, or the type of
advertising in fancy magazines, street billboards, the subway or a supermarket leaflet.
At all these levels the possible expression of ideologies is obvious, for
instance through the graphical emphasis of positive values with ingroups,
and negative values with outgroups. Through images, photos, text placement, page layout, letter type, colour and other graphical properties, thus,
meanings and mental models may be manipulated, and indirectly the
ideological opinions implied by them. A serious theory spells out what
graphical structures exactly may have which of these various functions?
202
Discourse
Sound
The phonetic and phonological expression structures of discourse (the
'sounds'), though systematically studied since the beginning of modem
linguistics and phonetics, have also been neglected in discourse analysis. 3
Articulation or auditory reception or phonemes may be marginal to a typical
discourse analyst who prefers to look at discourse structures beyond those of
words, phrases or sentences. Yet, pitch, volume and intonation are a rich
source of variation by which, as in graphical expressions, emphasis, prominence or distinctiveness may be controlled as a function of semantic and
ideological importance and relevance, as well as of opinion, emotion and
social position (as in authoritarian commands versus polite requests). Since
most conversation analysts work with transcripts, precisely these 'sound
structures' tend to be partly ignored in analyses, or reduced to rather crude
forms of representation or description, with the exception of the study of
applause in public address.
Especially interesting for ideological analysis is the fact that subtle sound
variation may directly code for underlying opinions in event and context
models, that is, without explicit semantic articulation: Admiration, praise,
derogation, blame and many other functions of discourse may thus be
signalled implicitly — and hence deniably — as a function of ideological
beliefs. The sound structures of talk to or among women and men, whites
and blacks, superiors and subordinates, and generally ingroup and outgroup
members, may thus display, emphasize, conceal or persuasively convey
ideologically based opinions about events or the participants in the context.
Morphology
The study of word-formation is not exactly a main focus of concern in most
types of discourse studies, and usually associated with traditional sentence
grammatical research. Since stylistic variation, compared with other levels
of utterances, is limited here, the ideological impact on the way words are
formed in text and talk seems to be marginal, especially in languages that do
not allow compounds. Where relevant, for instance in the study of neologisms, such ideological effects usually will be studied in lexical stylistics.
Syntax
On the other hand, the study of sentence forms, syntax, has drawn attention
from (critical) linguists interested in ideological analysis from the start. 5
Variatioñ in the order or hierarchical relations of the structures of clauses
and sentences is a well-known expression of dimensions of meaning as well
as of other underlying semantic and pragmatic functions. Thus, order and
hierarchical position may signal importance and relevance of meanings, and
Discourse structures
203
may thus play a role in emphasizing or concealing preferred or dispreferred
meanings, respectively.
Agency and responsibility for actions may similarly be emphasized or deemphasized, for example by active or passive sentences, explicit or implicit
subjects, as well as word order. It needs little analysis to show that such an
important function of syntactic variation may have an impact on the
description of ingroup and outgroup actions, and hence on ideological
implications of text and talk. Position and role of clauses may signal
implications and presuppositions, which are closely related to what language
users should or should not know, and hence to 6 the ideological discursive
functions of exposing or concealing information.
Among many other features of syntax, pronouns are perhaps the best
known grammatical category of the expression and manipulation of social
relations, status and power, and hence of underlying ideologies. Ingroup
membership, outgroup distancing and derogation, intergroup polarization,
politeness, formality and intimacy, and many other social functions may thus
be signalled by pronominal variation. Ideologically based respect to others
may be given or withheld by using familiar or polite pronouns of address, as
in French tu and vous and Spanish tu (or vos in some Latin American
countries) and Usted. Given the group-based nature of ideologies, group
polarization and social struggle is thus prototypically expressed in the wellknown pronominal pair of Us and Them. Indeed, there are few words in the
language that may be as socially and ideologically loaded' as a simple we.
The close relationship between group identity, identification and ideology,
as discussed before, explains this particular function of this pronoun.
The specific set of choices that are made among the possible structures of
syntactic forro in one specific discourse, is usually called the (syntactic) style
of that discourse. Combined with lexical variations in the choice of words
(lexical style, see below) such syntactic style is often studied in a separate
domain of discourse analysis, namely, stylistics. Style may generally be
described as the overall result of the consistent use of variable grammatical
structures as a function of properties of the context (or rather of the
interpretation of the context as represented in context models). This means
that style is by definition a function of the ideological control of such
context models, as we have seen for the example of polite or impolite uses of
forms of address.
Semantics
Graphics, sound and sentence forms are usually categorized as 'observable'
expressions of discourse, traditionally called 'surface structures' in generative grammar. In some critical and ideological studies (often but not only
in a Marxist tradition), such structures may even be called 'material',
although, as suggested before, there is very little 'material' in abstract
204
Discourse
structures (one reason I used 'observable' between quotes). Yet, in a
somewhat sloppy but practical sense, we may say that surface structures are
the kinds of things that are actually and 'observably' expressed, shown and
displayed for interpretation by recipients. But we should remember that also
these 'observable' structures of expression are in fact abstract or mental
structures being assigned, by theorists as well as language users, to the
various physiological, auditory or physical (phonetic, printed) properties of
communication.
The meaning of 'meaning'
Undoubtedly crucial in all ideological analyses of discourse are the meanings expressed by or assigned to surface structures by discourse participants.
Unfortunately, there are few notions in the study of language and discourse
that are so complex and vague as that of meaning. Especially also in critical
or ideological studies, the notion is sometimes used so broadly that it has
lost virtually all 'meaning'. That discourse expresses, conveys, has, constructs and does many other things with meaning is both commonsense and
scholarly knowledge. Yet, we need a sophisticated semantics, or even
various types of semantics, to be able to spell out how exactly, what kinds of
meanings, are involved here. Simply talking about the'production of
meaning', as is usual in much contemporary critical studies, does not tell us
much about the role of discourse or ideology in communication, interaction
and society.
Thus, it cañ t hurt to recall old linguistic, philosophical and logical
distinctions between (conceptual) meaning or intension, on the one hand,
and reference, that is, as a relation between expressions and things being
referred to, denoted or talked about (i.e. the referents, denotata or extension),
on the other hand. Similarly, in an abstract analysis, it also makes sense to
distinguish between word or sentence meanings, utterance meanings, speaker's meanings, hearer's meanings and socio-cultural meanings (including
ideological meanings).
As is the case for all structures of discourse, all these different 'meanings'
result from different theoretical approaches. In traditional linguistics as well
as in common sense, words are associated with (word) meanings, as is still
the case in dictionaries. In structural and generative grammars, meanings of
sentences are formally constructed as a function of the meanings of words
and syntactic structures. In philosophical logic, meanings are abstract
functions that make sentences true or false, or that pick out referents or
extensions (objects, properties, facts) in some situation or possible world.
Meaning and interpretation
In the philosophy of language as well as in psychology and most of the
social sciences, meanings are not so much abstract properties of words or
expressions, but rather the kinds of things language users assign to such
Discourse structures
205
expressions in processes of interpretation or understanding. This also allows
for contextual variation: a speaker and a hearer may assign (intend, interpret,
infer) different meanings to the same expression, and indeed, the same
expression may therefore also mean different things in different contexts.
Hence, meanings of discourse or language in use are contextual or situated,
and depend on the (interpretation of the) participants.
Psychologists will then further spell out how such meaning assignments
or interpretations take place mentally, and what memory representations
(such as models or knowledge) are involved in meaning production and
understanding. Socially oriented discourse analysis will usually ignore such
cognitive'processing' of meaning, and focos exclusively on the interactive
or social construction of meanings in or by discourse. In this case meanings
are usually inferred intuitively by the analyst, and are not further analysed. It
is on this (rather shaky) basis that much ideological meaning analysis often
takes place.
As we shall see later, discourse meanings are the result of selecting
relevant portions of mental models about events. That is, knowledge about
events is thus mapped on verbaily expressed meanings of text and talk, and
hence partly constrained by the possible word and sentence meanings in a
given language or culture. Since models embody opinions, which may in
turn have an ideological basis, also the meanings that derive from such
'ideological' (biased, etc.) models may embody ideological aspects.
Many of these opinions may be conventionalized and codified in the
lexicon, as the respective negative and positive meanings of the well-known
pair lerrorise versus 'freedom fighter' suggest. Lexical analysis is therefore
the most obvious (and still fruitful) component in ideological discourse
analysis. Simply spelling out all implications of the words being used in a
specific discourse and context often provides a vast array of ideological
meanings. As a practical method, substitution of one word by others
immediately shows the different semantic and often the ideological 'effects'
of such a substitution.
Theoretically, this means that variation of lexical items (that is, lexical
style) is a major means of ideological expression in discourse. Depending on
any contextual factor (age, gender, 'racé, class, position, status, power,
social relation, and so on) language users may choose different words to talk
about things, people, actions or events. Personal and group opinions, that is,
attitudes and ideologies, of participants are a prominent contextual constraint, and hence a major source of lexical variation. Given the obvious
ideological implications of lexical choice, we may also expect that language
users are often (made) aware of their style, and may hence also partly
control it, and thereby either emphasize or precisely conceal their 'real'
ideological opinions. The current debate on 'politically correct' language,
precisely focuses on this aspect of ideologically based lexical style, and
especially shows people' s positions in the relationships between dominant
and dominated groups.
206
Discourse
Propositions
Beyond lexical semantics, the study of discourse meaning of course has
many other aspects that are relevant for the mapping of ideology on text and
talk. Thus, first of all, the propositions that represent the meaning of clauses
and sentences have an internal structure, of which for instance the various
semantic roles (agent, patient, object, etc.) may exhibit the ways participants
are associated with an event, actively or passively, responsibly or as
experiencers of events and actions. In other words, semantic structures result
from model structures. Such semantic representations are obviously a
function of how events are interpreted and evaluated (in a model), and may
therefore be ideologically controlled, depending on the group membership,
the position or the perspective of the speech participants. Who is seen as the
hero or the villain, the perpetrator or the victim of an act, which roles need
to be emphasized or concealed, are questions that organize many ideological
attitudes, and such perceptions may directly be mapped into propositional
structures and their variable syntactic formulations (actives, passives, nominalizations and so on). $
Local and global coherence
Whereas most of the structures mentioned aboye are within the traditional
realm of Iinguistic grammars, discourse analysis was precisely developed in
order to account for structures and strategies beyond the sentence boundary.
Semantics (as well as pragmatics and interaction analysis) is especially well
suited to account for such more complex 'textual' meanings. Thus, sequences of sentences (or rather, of propositions) constitute discourses if they
satisfy a number of coherence conditions, such as (a) conditional relations
between the 'facts' denoted by these sentences, or (b) functional relations
(such as generalization, specification, contrast) among propositions.
Such coherence is based on the interpretation of events as represented in
the mental models of the language users, and may therefore also be
ideologically influenced. Whether language users see a social event as a
cause or not of another social event may thus have an effect on the
coherence of their discourse. In other words, coherence is both contextually
and socially relative, and depends on our ideologically controlled interpretation of the world.
The same is true for the kind of overall coherence represented by topics or
semantic macrostructures, which also signal what speakers or recipients
think is the most important information of a discourse. Such a judgement
may obviously be ideologically based. What for some is defined, topically,
as a 'race riot by a violent black mob', for others may be semantically
summarized as an 'act of urban resistance against racist police officers'. In
other words, semantic macrostructures (derived by special semantic 'reductioñ rules or strategies from propositions in models about an event) not only
define such important discourse structures as topics, overall coherence, or
Discourse structures
207
importance of information, but essentially also explain the well-known
ideological practice of 'defining the situatioñ .
The implicit and the explicit
Another ideologically relevant property of meaning is propositional relations, such as implication, entailment and presupposition. Thus, information
that is explicitly asserted may emphasize negative properties of outgroups or
positive ones about ingroups, whereas the reverse is true for implied or
presupposed meanings. The well-known ideological function of concealing
'real' social or political facts or conditions may be semantically managed by
various ways of leaving information implicit. This also shows the importance of distinguishing between mental modeis (beliefs) and discourse
meanings, although we often may infer what people 'really mean' (their
modeis) when they say something.
Similarly, we may describe acts or events in great detall, or do so only
with few details, or at higher levels of abstraction. Such variation may also
encode ideological positions — who, indeed, has interests in knowing or
concealing such details about social events? In sum, semantics is a rich field
of ideological'work' in discourse, and virtually all meaning structures are
able to 'signify' social positions, group perspective and interests in the
description of events, people and actions.
Schematic structures
Whereas topics represent the global meaning of discourse, overall schematic
structures or superstructures represent the global forro of text and talk. Such
global discourse forms or schemata are organized by a number of conventional categories, such as introduction and conclusion, opening and closing,
problem and solution, premises and conclusion, and so on. Stories, news
reports, conversations, meetings and scholarly articles, among many other
genres, are thus organized by conventional schemas that define the order and
hierarchical position of such categories (as well as the semantic macrostructures or topics that define the 'content' of these categories).
As is the case for the syntax of sentences, also this 'discourse syntax',
may vary and hence'code for' ideological positions. As is true for all formal
discourse structures, these schemata may signal importance, relevance or
prominence. What information appears in a headline, what is emphasized in
a conclusion, or what event descriptions count as complication or a
resolution of a story, depends on the ways events are interpreted, and hence
on ideologically variable positions. Obviously, some of these categories are
obligatory (as is the case for headlines of news reports), but others are not
(for instance background information in news reports), and also categories
may appear in different positions. Thus, greetings and leave-taking are
usually obligatory categories of conversation. Besides interactional functions, for example of address and politeness, they may also have ideological
208
Discourse
functions, such as when their absence is intended as an ideologically based
insult Similarly, if verbal reactions in a news report appear up front, we
know that the source of such reactions is found important, as are his or her
opinions, a structural feature that obviously has ideological implications. 9
Rhetorical structures
Discourse features a number of special structures or strategies that have been
amply described already'in classical rhetoric, and that are usually called
'figures of stylé , but which will here be called rhetorical structures. These
structures appear at all levels of discourse described aboye, and assign
special organization (repetition, deletion, substitution, etc.) to these levels,
for instance by the figures of rhyme and alliteration at the level of sounds,
parallelism at the level of syntax, and comparison, metaphor, irony, etc. at
the level of meaning. Unlike other discourse structures, these are optional,
and serve especially in persuasive contexts, and more generally to attract or
manage the attention of recipients.
In an ideological analysis this will usually mean that rhetorical structures
are studied as mearas to emphasize or de-emphasize meanings as a function
of ideological opinions. Metaphors may be chosen that highlight the
negative character of our enemies, comparisons in order to mitigate the
blame of our own people, and irony to challenge the negative models of our
opponents. Rhetoric, defined in this sense, is essentially geared towards the
persuasive communication of preferred models of social events, and thus
manages how recipients will understand and especially how they will
evaluate such events, for instance as a function of the interests of the
participants. It is therefore not surprising that rhetorical structures play such
an important role in ideological manipulation.'
Speech acts
Whereas utterances were traditionally analysed along two main dimensions,
namely, expressions (signifiants) and meanings (signifiés), the philosophy of
language and the social sciences have added an important third dimension:
action. Uttering words and sentences in text and talk, in a specific situation,
is also arad at the same time the accomplishment of a large number of social
actions, as well as participating in social interaction. Thus, assertions,
promises or threats are made, arad such speech acts are typically defined in
terms of social conditions of participants, namely, their mutual beliefs,
wants, intentions, evaluations and goals that have social implications.
Speech acts like threats presuppose power, and tell recipients that the
speaker will do something negative if they do not comply with his or her
wishes. Commands also presuppose power but require that the recipient
must do something. That is, relations between speech participants are crucial
in the ways speech acts are accomplished.
Discourse structures
209
This also means that if these social relations are ideologically grounded,
for instance in relations of dominante and inequality, such relations may
well also be displayed in the kind of speech acts speakers are (or feel
themselves) entitled to accomplish. At this point, the ideological control of
social practices directly impinges on speech acts, for example when whites
of equal social position feel entitled to give orders to a black person, or when
men threaten women. In sum, whenever relations between participants as
well as other dimensions of the context (time, location, etc.) are ideologically based, this may show up in the kind of speech acts being
accomplished by the participants.
Interaction
Finally, within the vast field of the social actions being accomplished in or
by discourse, we fmd a number of interaction strategies that express,
indicate, reflect or construct specific social relations between participants,
and which therefore are ideologically relevant. It is especially at this level of
analysis that social position, power and control of social members may be
exercised, opposed, mitigated or emphasized.
Interactional control may affect virtually all levels and dimensions of text
and talk. Powerful speakers may control context structures by requiring or
prohibiting the presence of specific participants, setting a time or place,
allowing specific gentes and not others, prescribing or proscribing the
language or professional jargon spoken, by initiating or changing preferred
or dispreferred topics or an agenda for a meeting, by sanctioning formal or
informal lexical style, by being polite or impolite, by (requiring) the
accomplishment of specific speech acts or the management of tums at
speaking, or by opening or closing the interaction, among many other ways
text and talk may be controlled. In all these forms of control, it is the social
position of the participants, and more generally the ideologically based
interpretation of the context that is thus being enacted, expressed or
constructed in talk.
More specifically, the interaction dimension of discourse is relevant in
everyday conversation and other forms of spoken, face-to-face dialogues
such as meetings and parliamentary debates. Such conversations are organized by a number of specific structures and strategies, for example those of
turn-taking, interruption or beginning and ending. Many of these are
obligatory and hence not directly controllable by ideologically variable
contextual factors. However, as is the case for interaction in general,
ideologically based group membership, power, positive self-presentation or
outgroup derogation are among the underlying social relationships that may
impinge on conversational structures and moves. That is, who may (or must)
begin or end the conversation or meeting, who may initiate or change topics
or who may interrupt whom, are among the many forms of power display in
discourse that may also have an ideological dimension, for example those
based on gender,'racé or class."
210
Discourse
Ideology and discourse control
In these and many other ways, thus, we see most concretely how relations of
dominance, conflict or competition between speech participants may implement and enact relations between groups. People not only engage in such
verbal social practices as individuals and as cultural members, but also as
members of specific groups, and such identities and membership may also
be locally negotiated. That is, group dominance is not simply mapped on
contextual relations between participants, but may be flexibly managed and
exercised in situationally variable ways.
The same is true for the ideologies that sustain such practices. From the
abstract level of group representations, they may provide particular opinions
about other group members which together with specific contextual constraints provide the unique interactional configurations we observe in
ongoing discourse. More generally, also for the levels we introduced aboye,
ideological mapping on discourse structures is seldom direct. It takes place
through more specific group knowledge and attitudes, the formation of
'biased' models of events and contexts, the construction of meaning representations, and the expression in variable forms and surface structures, in
ways that are a function of many social and contextuál constraints, of which
ideological beliefs are only one element.
For the practice of ideological analysis this also means that ideologies
cannot simply be 'read off text and talk. What is an ideologically relevant
expression in one discourse or context may not be one in another, or may
have an opposed ideological function at another moment. This means that
ideological discourse analysis is very complex, and needs to take into
account all levels of text and context, as well as the broader social
background of discourse and interaction. In the following chapters, I shall
discuss some of the topics of such an ideological discourse analysis.
Ideological discourse structures
265
events have a lower level of ideological expressions, implications and
functions, as may be the case for a TV instruction guide, an anide on
phonology, or a daily conversation on horticulture. That is, some genres
more typically function as persuasive expressions of opinions than others, if
only through the kind of topics that are associated with it. Most genres that
have persuasive functions or implications and that are on social topics have
ideological implications.
The second set of contextual constraint is the type of participant. Again,
people expect ideologically relevant social opinions from specific group
members rather than others. Thus, a politician, corporate manager, priest or
journalist writing or speaking about social issues is more likely to be (heard
as) expressing ideologically based opinions than a child or a carpenter
talking about how to make a table. Indeed, representatives of specific social
groups speaking about issues relevant to the group, for example wornen,
blacks, pacifists or environmentalists, will more typically be heard to express
ideologies than people who are not primarily speaking as group members.
This not only puts constraints on discourse structures, but also and importantly on the definition of the communicative situation by the recipient, that
is, on the recipient's context model, which will in turra monitor comprehension and event model formation.
That is, in many situations recipients already know that ideologically
based discourse may be expected from the speakers or writers. This implies
that ideological communication may be most effective when recipients do
not or hardly expect ideological implications, for instance in childreñ s
stories, textbooks or TV news, whose main functions are usually assumed
to be free of persuasive opinions. For news in most of the Western media
it is one of the major (ideological) criteria that'facts' should be separated
from 'opinioñ . It needs no comment that when such assertions are made,
that is, when ideology is denied, it is especially relevant to do ideological
analysis.
Besides types of communicative event and participants, there is another
context feature that is crucial in the reproduction of ideologies, namely,
properties of the intended recipients. That is, mass mediated or any other
kind of public discourse will have more serious ideological consequences, if
only because of the size of its audience, than mundane interpersonal
dialogues. Both genres may in specific contexts be equally ideological, but
ideologies expressed in public discourse convey opinions to many more
ingroup and outgroup members. Moreover, public discourse, such as that of
politics or the media, usually features institutional speakers or representatives who have more authority and hence more credibility. Much of the
ideological consensus construed among groups or in society today would be
difficult to obtain without coverage of relevant issues in the mass media.
This size of the audience of a discourse will be called its 'scope'. Trivially,
and all other things being equal, the larger the scope of a discourse, the
greater its ideological effects. And since those who have active access to,
and control over the mass media are generally members of the elites, larger
266
Discourse
scope will often be combined with higher credibility of the speakers/writers
and hence a higher chance that models will be construed as preferred.
Topics
Let us now turn to the question of what discourse structures are typically
involved in the expression or formation of ideology. 2
There are probably no structures of text and talk that have a more
prominent effect on the construction and further processing of models than
semantic macrostructures or topics. Derived (formally or by production and
comprehension strategies) from the propositions of a discourse or an event
model, they embody what is most relevant or important for the participante.
Unless recipients have alternative 'readings' of a discourse, topics will head
the model, and will generally be most accessible for further processing. If
people remember anything of a discourse at all after some delay, it is the
topic and maybe some details that are personally relevant for the recipient.
Since topics are represented by (macro) propositions, they may also
express opinions, and hence ideologies. These propositions may be
expressed in specific schematic categories of a text, for instance in the initial
summary of a story (of the type'What I particularly dislike of foreigners is
that they doñ t want to learn our language) or the headline of a news report
(131 ack youth involved in crime wave ). Ideologically based stereotypes and
prejudices may thus be highlighted twice: by their important semantic
function of a topic that organizes the semantic microstructures of a discourse, as well as by their schematic emphasis in the beginning or on top of
a story (often marked by special graphics, such as a banner headline, or by
special intonation in conversational dialogue). Obviously, the scope of the
newspaper report in Chis case, and hence the contribution to the reproduction
of racist ideologies in society, is vastly greater than that of an everyday story
among neighbours.
Since topics as expressed in discourse suggest preferred macrostructures
of event models and sine such macrostructures remain more accessible,
they also provide the'facts' that are used in the rhetorical arguments of
everyday conversation in support of ideological opinions ('Yesterday it was
in the newspaper that.
Similarly, these model structures will also be
used for further abstraction and generalization and hence as the basis for the
confirmation or construction of ideological attitudes and ideologies themselves, unless counter-information discredits the discourse or its writer/
speaker as being biased. In sum, discourse topics are crucial in the formation
and accessibility of preferred ideological models and, thus, indirectly in the
formation or confirmation of ideologies.
.').
Local rneaning
In discourse comprehension, prominently expressed topics play an important
role in the local comprehension of text and talk. They define the overall
Ideological discourse structures
267
coherence of the discourse. At the same time, they activate relevant
knowledge and help constnict the top level of the models being used for the
possibly biased interpretation of the rest of the discourse. Local meanings
may thus be ignored or fiterally'clown-graded' to the level of insignificant
detall.
Examining these local meanings as such, we deal with the actual 'contene
of discourse, and it is , here that most ideological beliefs will be incorporated
in text and talk. As we have seen before for the process of expression, this
means that beliefs in event models are selectively constructed to form the
semantic representation of text and talk. For obvious contextual reasons, not
ah we know about an event needs to be included in the meaning of a
discourse, so that speakers/writers make a selection, and it is this selection
that is fiable to multiple forros of ideological control. The general constraint
is contextual relevance: Those propositions are expressed which the speaker/
writer thinks the recipient should know. That such relevance decisions may
be in the interest of the speaker/writer is obvious; for instance, information
about an event that may give a bad impression of the speaker/writer, or
which in any other way may later be used 'against' the speaker/writer, may
be left out in order to influence the models of a recipient in the preferred
direction.
Here we encounter two important principies of ideological reproduction in
discourse, namely, the presence or absence of information in semantic
representation derived from event models, and the function of expression or
suppression of information in the interests of the speaker/writer. This last
principie is part of an overall strategy of ideological communication that
consists of the following main moves:
1 Express/emphasize information that is positive about Us.
2 Express/emphasize information that is negative about Them.
3 Suppress/de-emphasize information that is positive about Them.
4 Suppress/de-emphasize information that is negative about Us.
These four moves, which constitute what may be called the 'ideological
square', obviously play a role in the broader contextual strategy of positive
self-presentation or face-keeping and its outgroup corollary, 'negative otherpresentatioñ . Unlike the self-presentation moves usually discussed in the
literature, however, these are not primarily focused on participants as
individuals, but on participants acting as group members. This suggests a
third important principie of ideological discourse analysis, namely, the fact
that since ideologies are social and group-based, also the ideological
opinions expressed in discourse must have implications for groups or social
issues.
Detall and level of description
When applied to semantic analysis these principies and strategies allow a
wide variety of options. One was already suggested aboye — in descriptions
of situations (as represented in models of the speaker) some information may
268
Discourse
be expressed and other information may be left out. That is, in relation to the
original model, discourses may be relatively incomplete. If a news report of
a 'riot' only mentions the violence of 'a black mob' and not of the police, or
not the causes of the riot, then we typically have a description which is
incomplete relative to what is known and what would be relevant information about the 'riot'. The consequence of such relative incompleteness may
be incomplete models of recipients (e.g. the readers of the newspaper),
which may again have implications for the biased construction of attitudes,
as described earlier.
This semantic feature may also work in the opposite direction: discourses
may be relatively overcomplete when they express propositions that are in
fact contextually irrelevant for the comprehension of an event (that is, for
the construction of a model), but which are nevertheless included in the
semantic representation of a description. Following the moves of the
ideological square, we may assume that this will typically happen when such
overcomplete information negatively reflects back on outgroups (or positively on ourselves). The standard example in reporting on ethnic affairs is
to mention irrelevant ethnic group membership in crime reporting.
The same principies not only apply to the selection, inclusion or exclusion
of model propositions in the meaning of a discourse, but also to the level of
the propositions included. These may be quite general and abstract (as in
topics), but also very low-level and detailed. The ideological conditions and
consequences are the same. Biased discourses will tend to be very detailed
about Their bad acts and Our good acts, and quite abstract and general about
Their good acts and Ouf bad ones. Although the precise mental consequences of levels of description are not known, it seems plausible that
their results are more or less detailed models of events. Mentioning many
'preferred' details requires organization, that is, mapping on topics, so that
more or less detailed text fragments nevertheless get topical status. This will
in turra allow them to be recalled better than a description of the same
sequence of events with just one global proposition. This is especially also
the case when details are 'vivid', for example when much 'visual' detall is
presented of actions. Precisely such details may imply (unstated) negative
evaluations which in turra may be taken up in the topical proposition
summarizing this event in the model of the recipient. Although these and
many other assumptions of this theoretical analysis of ideological discourse
structures need to be empirically tested, they are consistent with what we
now know about discourse processing. 3
Implicitness versus explicitness
The well-known semantic properties of implicitness and explicitness of
discourse can easily be explained in terms of mental models — implicit
information is the information of a mental model that could or should have
been included in the semantic representation of a discourse. As is the case
for the level of specificity and the relative in- or overcompleteness of
Ideological discourse structures
269
descriptions, we may more generally say that propositions may selectively
be made explicit or left implicit as a function of the interests of speakers as
group members. Besides relevant components of actions, this may typically
be the case in the expression of conditions (causes) and consequences of
events, as was suggested for the frequent omission of causes of ethnic
conflict that negatively reflect on our ingroup (e.g. police brutality, inner city
neglect, poverty, unemployment or discrimination by employers). On the
other hand, ideologically blaming the victim in this case means that the
negative properties attributed to the outgroup (e.g. drug abuse, cultural
deviance) will be made explicit. Research on the representation of ethnic
affairs in the media has often found such ideological dimensions of semantic
implicitness or explicitness. 4 (Por detailed examples, see Chapter 28.)
One step between presence and absence of information is when propositions are not as such expressed in discourse but implied by other propositions that are expressed. Implication and presupposition are the familiar
semantic relations involved here, and both involve inferences based on
models and social knowledge. The ideological function of the use of such
semantic relations is not always straightforward. Following the ideological
square, we may assume, as aboye, that implied information is not explicitly
asserted, and hence not emphasized, and will therefore typically be information that needs to be concealed in the interest of the speaker and the ingroup.
This is especially so when the implied information cannot be readily inferred
from socially shared knowledge. When such implied information needs to be
known in order for propositions in the text to be true or false, we speak of
presuppositions, and these may have the same ideological functions —
information is assumed to be 'given' or 'true' and is therefore presupposed
by the discourse, but it may well be that the presupposed information is
questionable or not true at all. That is, in this case it is obliquely asserted to
be true, but without emphasizing such an'assertioñ . Following the strategies of the ideological square, it is easy to spell out what information about
ingroups and outgroups will typically be expressed and which information
will be left implicit.
Local coherence
S equences of propositions are linearly connected by relations of 'local'
coherence. Such conditions of coherence are first of all defined relative to
the event models. Two propositions are coherently related if they express
'facts' in a mental model that are (e.g. causally, conditionally) related. But if
mental models are ideologically biased, this also means that discourse
coherence may be biased and have biased models of recipients as a result.
Taking the same example of a 'luce riot' discussed aboye, a police report
whose version of the facts, that is whose underlying model, is adopted by the
press, may describe the events such that criminal behaviour of black youths
is taken as the cause of the riot, and not 'tough' policing. Similarly, coherent
explanations of social events are in general based on assumptions about
270
Discourse
causes and consequences, such that the ideological bias of coherence may
presuppose or imply biased models of the social situation.
Propositions may also be related in a sequence by means of 'functionar
semantic relations, such as generalization, specification, example or contrast.
In ideological discourse these play an important role because they manage
the way statements are understood in relation to other ones. Por instance, a
prejudiced story about minorities may feature descriptions of negative events
about minorities, followed by the generalization, 'They always do that.'
Such a generalization is of course crucial in the transition from models to
generalized models and social representations. It persuasively suggests that
this was not merely an incident or a personal experience, but a general,
structural phenomenon. In this way, concrete events (and their models) are
related to, and at the same time explained and legitimated by general
attitudes.
Also the converse takes place: a speaker may make a general, prejudiced
statement about immigrants, and knowing that such a generalization might
be understood as prejudiced, may then add 'evidence' in the form of an
example, specification or a whole story. Similarly, group polarization may
be discursively emphasized by typical semantic and rhetorical contrasts, as
in, 'We always have to work hard, and they only have to ask for welfare.' We
fmd another well-known type of contrast in disclaimers such as 'I have
nothing against Turks, but. . .', in which something positive about Me (Us)
is being combined with a negative statement about Them. That is, such
disclaimers also play a role in the complex strategies of positive selfpresentation and negative other-presentation which is so typical for ideological discourse.
Lexicalization
The most obvious and therefore most widely studied form of ideological
expression in discourse may be found in the words being chosen to express
a concept. The pair 'freedom fighter' versus lerrorise is the paradigmatic
example of this kind of ideologically based lexicalization. That is, a negative
concept of a group is represented in a model, and depending on context, the
most 'appropriaté word is selected, in such a way that an outgroup is
referred to and at the same time an opinion about them.
Following the ideological square, this means that in general we may
expect that, depending on context, outgroups will be described in neutral or
negative words, and ingroups in neutral or positive tercos. And conversely,
we may also expect that in order to describe groups and their practices,
various forms of mitigation and euphemisms may be selected, thus adding a
rhetorical dimension to lexicalization.
Finally, lexicalization may also extend to the nominalization of propositions, such that agents or patients are left implicit. Inner city 'policing' thus
focuses on a verb, without actually making explicit who is being policed,
whereas the role of the police is also de-emphasized. It need not be repeated
Ideological discourse structures
271
what influence such nominalizations may have on the structuring of action
roles in the models of recipients.
Discourse schemata
Discourses not only have a global meaning but also a global form or
conventional schema, which consists of a number of characteristic categories
appearing in a specific order. Thus, arguments may feature various kinds of
premises and a conclusion; stories are organized by narrative schemata with
such categories as orientation, complication and resolution; and news reports
begin with the well-known category of a Summary consisting of a headline
and lead. As is the case for global meanings or topics, also such schemata
function as organizers for complex information, and at the same time as
properties that help define discourse genres. Stories organized by a conventional schema are thus easier to tell, understand and memorize, while a
headline in a news report has the conventional function of expressing the
main topic, so that readers know what the report is about and may decide to
read it or not.
Since these categories are conventional, and vary between genres and
cultures, they also have important social functions. Making headlines for a
news report is part of the routines of newsmaking, and so is finding quotes
for a verbal reaction category in a news report. As is the case for the
organization of everyday conversations (beginning with greetings and ending with leavetaking) or the schematic organization of meetings, sessions
and other institutional communicative events, these schemata organize
discourse as much as they do interaction.
Given the important cognitive and social functions of schemata, it stands
to reason to assume that they may also have ideological functions. It is vital
whether information is being expressed in a headline or not, and this may of
course influence the forro of resulting models; negative information or
opinions about minorities may thus appear in the headline, and information
that is important but positive about them may be excluded, as much research
on 'ethnic' news shows. The same is true for the appearance of opinions in
conclusions of arguments, which social groups have 'access' to the verbal
reactions category of a news item, and so on. Information and opinions
about Us and Them may be further organized, and be made more or less
prominent through such schemata.
Style
Lexicalization may vary as a function of opinions, and if such takes place
throughout a discourse, we would then speak of a specific lexical style.
Generally, then, given specific meanings or model information, different
expressions may be used to express such 'content', and this variation may
signal in many ways the social context of the communicative event.
272
Discourse
Depending on the nature of the communicative event, the genre, setting or
participants, thus, 'surface' structures (lexical items, syntactic structures,
pronunciation and graphics) may vary in order to intentionally or unintentionally signal their contextual boundedness: the situation may be more or
less formal, the relations between the participants may be friendly, familiar
or distant, and participants may have various opinions about each other. The
result may be a more or less formal, familiar or polite style, and at the same
time an indication of underlying ideological 'stances' of speakers. Everyday
racist events, for instance, frequently involve 'breaches' of appropriate
interactional style, for instance when white speakers use derogatory words or
impolite pronouns to or about minorities. 7
In sum, lexical and grammatical style is one of the most obvious means
speakers have to explicitly express or subtly signal their ideological opinions
about events, people and participants. The same is also true for syntactic
structures and their possible variation. Sentences may be expressed in an
active or passive voice, and agents and patients of actions being described by
such sentences may in this way be made more or less prominent or
completely left implicit, as is the case in nominalizations, as suggested
above. 8 More generally, word order, clause structure or clause relations may
put information in more or less prominent positions, and as is the case for all
structures and strategies discussed here, this will subtly effect processing and
the construction of models. According to the ideological square, we will thus
fmd that positive action roles of outgroup members will be put in less
prominent order or position, and vice versa for their negative action roles
(and conversely for the positive and negative roles for ingroup members).
Style thus may signal in many ways the structures of the social context,
including relationships of power. A powerful social position of a speaker
will thus not only be 'expressed' by the words or syntax being chosen, but is
at the same time enacted and reproduced by it. This may become apparent in
stylistic differences between male and female talk and text, as well as that
between majorities and minorities, doctors and patients, civil servants and
clients, professors and students, judges and defendants, or police officers and
suspects. Style thus defines positions of participants, and wherever these are
controlled by ideologies, as is the case with the examples just mentioned,
style will be a direct'tracé of ideologies in discourse. Social discrimination
is thus implemented directly by those who control the style of text and
talle.
Rhetoric
Several examples have already been given aboye of the rhetorical dimensions of discourse, defined here (rather narrowly) as the system of special
'rhetorical figures' that have specific persuasive functions at various structural levels of discourse, such as metaphors, euphemisms, irony or contrasts
at the semantic level, or alliteration and rhyme at the phonological level.
Ideological discourse structures
273
Similar observations may be made for graphical structures, which are mainly
organized to control attention and steer interpretation through emphasis.
The main function of such rhetorical structures and strategies is to manage
the comprehension processes of the recipient, and hence, indirectly the
structures of mental models. A specific negative opinion may be emphasized
by a catchy metaphor from a negative conceptual domain (for instance,
describing outgroup members in tercos of animals such as rats, dogs,
bloodhounds, snakes or cockroaches), by comparisons of the same type, or
by hyperboles describing their negative characteristics. Repetition moves
such as syntactic parallelism, rhyme or alliterations may further increase the
attention paid to such semantic properties of the discourse, and thereby
enhance the possibility that they will be stored, as intended, in the preferred
model of an event. The converse is true for negative properties of ingroup
members, in which case we will expect various forms of rhetorical mitigation, such as euphemisms, understatements and other ways to deflect
attention from specific meanings.
Interaction strategies
Finally, and specifically for spoken dialogues, many of the structures
discussed aboye will be further accompanied by moves and strategies of an
interactional nature. If the basic aim of ideological communication is to
influence the models and social representations of recipients in such a way
that preferred opinions are prominently represented, recalled and eventually
accepted, also several forms of interaction management will play a role in
this form of social 'mind control'.
First of all, however, it should be stressed that interactional strategies
themselves are hable to ideological control, as is also the case for the context
and its models. Ideologically based dominance and inequality is not only
expressed in the structures of text and talk discussed aboye, but also in group
relations as embodied in participant roles and actions. In the same way as
speakers may control topic or style, they may control turn taking, 'schematic' sequences (who begins or closes a dialogue, meeting or session), pauses,
laughing and so on. Power abuse by speakers of dominant groups may thus
also be blatantly or subtly enacted by limiting the conversational freedom of
others. If women, minorities, students, clients, patients or 'ordinary people'
have less to 'say' in society, this will also show and be reproduced in many
conversational situations. Detailed conversation analysis has shown how
such forms of social inequality may be enacted in the subtle details of
mundane and institutional talk and interaction. 9
At the same time, these interactional strategies may have an effect also
during the construction of (semantic) event models. This is obvious for the
interactional control of meaning, for instance in topic management, as
described aboye. However, control of interaction itself, such as in turn
taking and sequencing, may also influence the ways recipients construe
274
Discourse
models of events. For instance, participant roles are important in ideological
communication, for instance in the management of credibility. Power and
status of speakers is a well-known condition in the way assertions are
accepted by recipients. However, conversation analytical research would
correctly observe that such social properties are not simply something
people 'have', but that they are (also) interactional accomplishments. Status
and power are contextually enacted and thus reproduced in many subtle
ways, such as bodily position, distance between speakers, clothing and
props, and the ways speakers control talk.
Crucially then, models and their representations depend on who says
what, and interactional management may control such effects. For instance,
speakers may be prevented from saying dispreferred things by interruptions,
or alternatively be encouraged to speak by selective turn allocation if they
are expected to say preferred things. Similarly, interactional strategies of
displaying agreement and disagreement play an important role in the
management of event models and their opinions. Specific speech acts
(commands, orders) may be enacted to implement social power, but also to
emphasize the negative characteristics of outgroup members (accusations,
blaming the victim). These are merely some of the many examples of the
ways interactional moves and strategies express, implement, enact or accomplish ideologically based opinions, perspectives and stances of speakers, and
the ways the models of recipients are shaped according to the preferences or
interests of speakers or the groups or organizations they represent.
Manipulation
Ideological communication is often associated with various forms of manipulation, with strategies that manage or control the mind of the public at
large, and with attempts to thus manufacture the consent or fabricate a
consensus in the interests of those in power. 1. Indeed, modem power and
ideological hegemony are precisely defined in tercos of effective strategies in
the accomplishment of compliance and consent, so that people will act as
desired out of their own free will. In that case, power and dominance will
seem natural, legitimate and commonsensical, and will be taken for granted
without significant opposition.
Formulated in this way, we get a simplified picture of the complex
processes at work in the enactment of dominance and the accomplishment of
hegemony. Without a much more detailed study of the social, cognitive and
discursive elements of the structures, strategies, processes or representations
involved in this forro of the 'modem' reproduction of dominance and
ideologies, such analyses barely go beyond easy slogans or superficial social
analysis and critique.
In the previous chapters and aboye, I have outlined some ideas about the
mental structures, social conditions and discursive reproduction involved in
the reproduction of dominance and hegemony. A study of manipulation,
Ideological discourse structures
275
mind control or the manufacture of consent needs to take place in such a
complex framework." Aboye, I have given some examples of how ideologies are expressed and especially persuasively conveyed by text and talk,
and how models and social representations may be effected by the structures
of discourse and context.
Thus, manipulation basically involves forms of mental control of which
recipients are not or barely aware, or of which they cannot easily control the
consequences. Models are constructed of events in a way that has implications for the construction of shared social representations people have about
the world, which in turn influence the development or change of ideologies.
Given the fundamental role of ideologies in the management of social
cognitions and models for discourse and other social practices, ideological
control and compliance are the ultimate goal of hegemony. We have seen
how specific discourse structures and strategies, such as the control of topics,
style or interaction strategies, may have such influences on models and other
representations of the mind. Because of such discursive properties, knowledge about events will he incomplete or biased in favour of speakers or their
ingroup, and this may affect more general knowledge about the world. Even
more crucially, this is the case for the management of opinions, in such a
way that a negative opinion about specific outgroups seems the most
'natural' or conclusion from the models as persuasively controlled
by discourse.
Conclusion
Of the vast richness of discourse structures and strategies I have mentioned
only a few. A detailed study will be necessary to atan all possible ways in
which contextualized text and talk exhibas and reproduces ideologies.
However brief, the discussion shows the basic principles at work. Ideological communication is a double-sided process, in which ideologically based
beliefs are expressed (or concealed), and persuasively control the minds of
recipients. Mind control is obviously an exceedingly complex process. But
also here, some basic formats of ideological influence seem to emerge from
the analysis — in order to contribute to the construction of preferred models
in a given context, discourse structures must be shaped in such a way that
specific model structures are the most likely consequence.
In the ideological situation of dominance, power abuse, group conflict or
competition, this in general means that (members of) outgroups need to be
treated and portrayed negatively, and (members of) ingroups positively. This
principie applies both to the pragmatic or interactional context, as well as to
the forms and meanings of text and talk. At each level of analysis, thus, we
find emphasis (prominence, importance, focus, etc.) on our good things and
their bad things, and vice versa for our bad things and their good things.
Besides this control of group-related opinions about Us and Them and their
properties and actions, discourse structures more generally control the
276
Discourse
management of the structures of models and social representations, for
example through explicitness versus implicitness, manifestation versus concealment, levels or details of description, the distribution of agency, responsibility or blame, the relationships between facts, and so on.
In sum, whatever the ideological shape of underlying attitudes, they will
appear in models of speakers, and these will try to appropriately and
effectively express such social representations in text and talk and their
contexts, in a way that most likely results in the construction of preferred
models. Often, and especially in what we call manipulation, this happens
without the awareness of recipients. It is more or less in this way that
ideologies are reproduced in everyday life. Later studies of discourse and
ideology will have to spell out the details of the general framework
presented here.
The Ideology and Discourse of
Modem Racism
A concrete example
After the theoretical chapters of this book, let me finally analyse a concrete
example. In fine with my choice of racism and racist ideologies as illustration of general principies, this chapter examines in some detall the ideology
and discourse as expressed in a recent book: The End of Racism: Principies
for a Multiracial Society, by Dinesh D'Souza (New York: Free Press, 1995).
Also in some of his other books, for example on multiculturalism, D'Souza
has made himself a vociferous spokesman of the New Right in the USA, and
a staunch defender of conservative ideas. Indeed, we might call D'Souza one
of the main'ideologues' of contemporary conservative ideologies in the
USA.
In the End of Racism D'Souza deals with what he sees as a 'civilizational
crisis' in the USA, and focuses on what he consistently calls the 'pathologies' which, according to him, characterize the African American community in general, and the black'underclass' in particular (in my analysis,
words in my running text actually used by D'Souza will be indicated by
double quotation marks). Given the size of this book (724 pages), this is no
mere ideological tract. On the contrary, D'Souza has set himself the task of
writing a broadly documented study of the ethnic and racial situation in the
USA. An endorsement by George M. Frederickson in The New York Review
of Books, printed on the cover, says: 'The most thorough, intelligent, and
well-informed presentation of the case against liberal race policies that has
yet appeared.'
Thus, D'Souza deals with what he sees as the breakdown of the 'liberal
hopé of race relations in the USA, the origins of racism, slavery, the rise of
liberal anti-racism, the civil rights movement, Eurocentrism and Afrocentrism, the IQ debate, fmally culminating in an apocalyptic vision of the
'pathologies' of black culture. In many respects, this book may be seen as
the ideological foundation of a conservative programme of race relations in
the USA. Since D'Souza is a scholar attached to the conservative think tank
of the American Enterprise Institute, we may conclude that his book does
not merely express a personal opinion, but also has a powerful institutional
backing. We already saw in Chapter 19 that contemporary ideologies are
often produced and reproduced by such ideological institutions.
278
Discourse
Given his right-wing radicalism in ethnic—racial matters, D'Souza has
been severely criticized, and accused of racism (in the introduction to the
second edition of the book he discusses and rejects such critique). After
having examined his theses and evidence in detall, and analysed the
discursive formulation of his underlying ideologies, I have come to the
conclusion, with others, that this book indeed articulates a special form of
'cultural racism', celebrating white, Western cultural and civilizational
hegemony, and especially problematizing and attacking African-American
culture. As is also clear from much of the literature on'modem racism',
most forms of racism are no longer biologically based, but take a more
'acceptablé form as cultural racism: others are not vilified for what they are,
but for what they do and think. More generally, D'Souza defends ideas that
are sometimes called 'symbolic racism': a forceful rejection of any form of
affirmative action, a strong repudiation of egalitarian values, problematization of blacks, blaming the victim, and so on. 1 Indeed, he even proposes
the repeal of the Civil Rights Act of 1964 (p. 544), and he favours 'rational
discrimination' in the prívate sphere.
Our ideological and social enemies and Us
Given their multiple group memberships, individuals may acquire and
personally adapt several ideologies or ideology fragments. This means that
D'Souzá s book is not merely an expression of conservatism and modem
racism, but a personal combination of these and other ideologies, attitudes,
beliefs, values, models and other social and personal representations.
Yet, where he expresses positions and opinions that seem to be widely
shared, at least among conservatives, in the USA (and also in Europe), we
may assume that he is not merely writing as an individual, but also as a
member of several ideological communities. At the end of his book, thus, he
explicitly aligns himself with other 'cultural conservatives' (p. 521). His
opinions about multiculturalism, affirmative action, the inner city ghettos
and related topics are widely shared by other conservatives in the USA.
Hence, abstracting from more personal views, we may read and analyse his
book as a formulation of group ideologies. 2
The ideological enemy
Ideologies are often formulated, explicitly or implicitly, as attacks against
ideological opponents or enemies. Anti-Communism has been the most
prominent example, especially in the USA, of this kind of anti-ideology. In
D'Souzá s book, this ideological enemy is what he calls 'cultural relativism', whose major tenet is that all cultures are equal, and that we should not
assume any value hierarchy between different cultures. D'Souza traces this
tendency to early twentieth century anthropology, and especially to Franz
Boas and his students.
The ideology and discourse of modem racism
279
Throughout his book, cultural relativism is frequently blamed for virtually
all ills of US society, and especially as the ideological source of contemporary 'anti-racise policies and practices in the USA:
[1] [The main problem is] Liberal anti-racism. By asserting the equality of all
cultures, cultural relativism prevents liberals from dealing with the t ation's
contemporary crisis — a civilizational breakdown that affects all groups, but is
especially concentrated among the black underclass. (p. 24)
[2] Fundamental liberal principies are being sacrificed at the altar of cultural
relativism. In its fanatical commitment to the relativist ideology of group equality,
liberalism is inexorably destroying itself. (p. 530).
[3] Relativism has become a kind of virus, attacking the immune systems of
institutional legitimacy and public decency. (p. 532)
As these examples also show, the reference to liberalism as an ideological
orientation is at least ambiguous. On the one hand, the specific US sense of
politically or culturally 'progressive' may be meant by it (as in example 1,
whereas D'Souza himself does not deny his allegiance to the original,
philosophical-political meaning of the term, as in example 2. We may
therefore expect, as was argued in the previous chapters, that the ideological
conflict presented in his book will be articulated in starkly polarized terms,
where all They think is inherently bad, and all We think is inherently good.
The rhetoric and lexical style of these examples expresses this ideological
polarization, as is shown in the use of metaphors from the domain of health
('virus', Immune system') in 3 and from traditional religion ('sacrificed on
the altar of`), as well as by the use of hyperboles ('civilizational breakdown') in example 1. The rhetorical contrast in 3 suggests that there is a
struggle between Us and Tbem. They are enemies who 'attack' us, and We
defend — as an 'itnmune system' — legitimacy and decency in the USA.
Framed in those terms, the ideological debate turras into a fierce struggle
between Good and Evil, as was also the case in classical anti-communism
until the Reagan era.
The social enemy
D'Souza and his fellow cultural conservatives not only have an intellectual,
ideological enemy, but also a social one, namely, African-Americans.
Although, as we shall see in more detail below, he emphasizes that his
animosity is not directed against blacks as a'racé, but rather against
African-American culture, his special focus on blacks can hardly hide the
fact that he is not merely fighting a cultural war. It is this reason why in his
book, and its underlying ideology, 'culture' and 'ethnicity' represent the
respectable mask behind which (acknowledged) ethnocentrism mingles with
various brands of modem racism. Although much of his fury targets the
black 'underclass' and its social 'pathologies', he often forgets this specification and problematizes the whole black 'culture', which he sees as
coherent and associated with all African-Americans in the USA.
280
Discourse
This is a very anti-black book. If D'Souza had more generally been
worried by the 'breakdown of civilization', as he so hyperbolically calls the
present 'crisis' in the USA, he could have targeted many other social or
cultural groups. With many of the same argumenta and examples, he could
also have focused on Latinos, on Native Americans, on the 'dependen
white underclass, on all unmarried mothers, all criminals, or all minorities
who profit of affirmative action. He does not. He specifically singles out
blacks, and his extremely biased, if not racist, judgements barely leave
another conclusion than that these are his real social enemies:
[4]The last few decades have witnessed nothing less than a breakdown of
civilization within the African-American community. The breakdown is characterized by extremely high rates of criminal activity, by the normalization of
illegttimacy, by the predo minan ce of single-parent families, by high levels of
addiction to alcohol and drugs, by a parasitic reliance on government provision,
by a hostility to academic achievement, and by a scarcity of independent
enterprises. (p. 477)
This quote sums up D'Souza's major points of resentment against the
African-American community. Indeed, he does not speak here of a (relatively small) section of this community, but of the community as a whole.
Where many others would talk of 'social problems' of some inner-city areas,
D'Souza's view is more apocalyptic. He sees 'nothing less than a breakdown
of civilization'. In many places of his book, he explicitly speaks of AfricanAmericans as a 'threat' not only to themselves but to the whole society:
[5]The conspicuous pathologies of blacks are the product of catastrophic cultural
change that poses a threat both to the African-American community and to society
as a whole. (p. 478)
Whereas conservatives before had communists as the major internal as
well as external enemy, this kind of socio-political paranoia now targets
blacks. In order to emphasize the 'pathologies' of blacks, the Asian
community in the USA is held up as the good example, an example that at
the same time serves as a strategic argument against those who might see
racism in D'Souza's attacks against blacks:
[6]By proving that upward mobiliry and social acceptance do not depend on the
absence of racially distinguishing features, Asians Nave unwittingly yet powerfully challenged the attribution of minority failure to discrimination by the
majority. Many liberals are having trouble providing a full answer to the awkward
question: 'Why can't an African-American be more like an Asian?'
One might easily explain this racial divide-and-tale principie by the fact
that D'Souza himself is an example of the Asian success story (he is from
India), but there are few other traces of his Asian (or Indian) allegiances in
the book. He does not speak for immigrants or minorities at all. On the
contrary, as is true of many conservative immigrants, he completely identifies with Western civilization, and the dominant white majority which,
obviously, could not have a more persuasive spokesman when it comes to
attacking multiculturalism and affirmative action: who is more credible in
attacking the others than one of them? As may be expected, conservative
The ideology and discourse of modern racism
281
blacks and other people of colour in the USA are extensively celebrated and
promoted, and have foil access to the media and other ideological institutions, especially when they serve as 'useful idiots' and sustain the dominant
consensus of the white elites.
Obviously, such groups and group relations need to be located in the more
complex, socio-political and intellectual framework of US society. Thus,
among the ideological enemies (the 'relativista' or the 'Boasians' — from
Franz Boas, famous US anthropologist) he further identifies most liberal
(progressive) scholars, politicians and joumafists, proponents of civil rights
and affirmative action, anti-racists, and al' those whom he portrays as
condoning or having vested interests in the continuation of 'black pathologies'. One stylistic ploy in the derogation of his ideological enemies is to
call them 'activists', including professors whose opinions he dislikes. In
passing he also includes some other target groups and ideologies of
conservative scorn:
[7] activists draw heavily on leftist movements such as Marxism, deconstructionism, and anticolonial or Third World scholarship. (p. 345)
[8] ... solutions [of African-American scholar Cornell West] are a quixotic
combination of watered-down Marxism, radical feminism, and homosexual rights
advocacy, none of which offers any realistic hope for ameliorating black pathologies. (p. 520)
In sum, although not the main target of his ire, his ideological enemies
stretch far along the social horizon, and include all progressive, altemative
or otherwise non-mainstream groups and the institutions associated with
them.
Us
Whereas there is little ambiguity about who bis enemies are, who are We in
this polarized representation of the civilizational conflict? As usual in this
kind of discourse, We are largely implicit and presupposed, and in need of
much less identification. In a large part of this book on the 'breakdown of
civilization', We are simply all civilized people. More specifically, also in
the historical sections of the book, We are those (mostly Europeans) who
invented'Westerñ civilization. Within the context of the USA, We may
variously be all non-blacks, or whites, or all those opposed to multiculturalism, affirmative action and state interference.
Whereas his positive descriptions of all these different We-groups with
which D'Souza identifies leave no doubt about his allegiances, his closest
ideological reference group comprises what he calls the 'cultural conservatives':
[9]The only people who are seriously confronting black cultural deficiencies and
offering constructive proposals for dealing with them are members of a group we
can call the reformers. Many of them are conservatives... (p. 521)
282
Discourse
They are the ones who, at the end of the book, have 'understood' the
seriousness of the 'civilizational breakdown' in the African-American community, and have made proposals to amend it. Quite predictably, D'Souza
includes a group of conservative blacks among their ranks, and does not
seem fazed by the inconsistencies such a selection engenders when he at the
same time lambasts the entice African-American community Apparently,
and as always, there are exceptions, and those are Our friends.
Since ideologies articulate within and between groups, we now have the
first elements of the social framework that sustains D'Souza's ideologies.
We know his enemies and we know his friends, and we know that he serves
as the ideologue for these friends, and as the ideal opponent of his
enemies.
The conflict and the 'crisis'
Ideological struggles are rooted in real political, social or economic conflicts. They do not merely involve arbitrary groups, but involve group
relations of power, dominance or competition. At stake is access to scarce
social resources, both material as well as symbolic ones. The conflict that
serves as the background for the ideological struggle in which D'Souza takes
part involves both 'race' and class, and especially focuses on the relations
between the white majority and the African-American minority in the
USA.
As is also obvious from the historical chapters of his book, this conflict
has a long history: European world exploration and colonization, the
enslavement of Africans by Europeans (and Arabs), the plantation economy
in the rural South, abolition, the emergence of scientific racism, the Jim
Crów laws, racial segregation, the civil rights movement, the end of formal
segregation and official racism, affirmative action, large scale immigration
from Asia and Latin America, multiculturalism in education, and finally the
conservative backlash of which D'Souza's book is a salient example.
Despite their 'real' socio-economic backgrounds, conflicts are sociopolitical constructs, which are defined differently by the various groups
involved in them, depending on their ideological orientation, group goals
and interests, as well as the everyday experiences of their members. Ongoing
sociopolitical conflicts such as that of race relations in the USA are
characterized not only by the many structural properties of social inequality
and occasional reform. They also know a series of 'crises', which are also
defined by shared mental representations of (and hence differently interpreted by) groups in conflict. A crisis may occur when one of the participant
groups enhances its political, economic or ideological dominance and
oppression or when the dominated group engages in explicit forms of
resistance. Thus the conservative backlash that coincided with conservative
Reagonomics and the victory of neo-liberalism in the 1980s and 1990s, is
one of such crises. This crisis in turn found its ideological motivation in the
The ideology and discourse of modern racism
283
reaction against the (modest) political and economic gains of AfricanAmericans that resulted from another crisis, namely, the civil rights movement and the social govemment policies of the 1960s and 1970s.
The social and political function of D'Souza's book should be defined
against this general background of race relations, politics and policies in the
USA, but draws its rhetorical relevance and persuasiveness especially from a
self-defined 'civilizational crisis'. That is, structural properties of US society
(such as poverty, especially in the black ghettos or overall socio-cultural
changes) are interpreted and presented as a major threat. Once defined as
'catastrophic', such a perceived threat demands urgent action and policy,
and D'Souza's book provides the ideological principies for such a 'multiracial society', as its subtitle specifies. We have seen that just talking of
(well-known) 'problems' will not do in such a rhetorical book. Hence such
social problems need to be magnified to a disaster of major proportions, as
also the frequently hyperbolic style of D'Souza shows:
[10]. . . the natioñ s contemporary crisis — a civilizational breakdown that affects
all groups ... (p.24)
[11]... a deterioration of basic civilizational norms in the ghetto. (p. 241)
[12]The conspicuous pathologies of blacks are the product of catastrophic
cultural change that poses a threat both to the African-American community and
to society as a whole. (p. 478)
[13]For many whites the criminal and irresponsible black undérclass represents a
revival of barbarism in the midst of Western civilization. (p. 527)
In other words, we do not merely have a conflict between two groups,
whites and blacks, in the USA, but a momentous struggle, namely, that
between (white) 'civilization' and (black) 'barbarism'. And, as may be
expected, D'Souza is the hero who has taken on the Herculean task of
fighting the forces of barbarism, as also the Greek heroes defended their
civilizations against the barbarian foreigners. D'Souza explicitly refers to the
Greek history of 'Western civilization' and democracy, as an example
which, until today, deserves emulation, including 'rationar, ethnocentric
discrimination of the barbarian others. Thus, his struggle is not just one that
tries to safeguard the interests and privileges of the dominant, white middle
class, but more grandly presents itself as a valiant defence of Western
civilization against the onslaught of a 'rainbow' coalition of blacks, immigrants, leftists, gays, lesbians, multiculturalists, Boasian relativists and
others who threaten the status quo. In that respect, D'Souza and his book,
and the ideologies he defines, are quite coherently conservative and ethnocentric. Let us now try to reconstruct these ideologies and other social
representations from his book and then examine in some more detall their
persuasive discursive manifestations.
284
Discourse
Reconstructing ideologies
Recall that ideologies, defined as basic social representations of groups,
should not be identified with their discursive expression. Indeed, the relation
between ideologies and discourse may be very indirect — usually, more
specific beliefs from social attitudes and from personal models of events
show up in text and talk, further modified by the constraints of context
models of speakers and writers. That is, more often than not, ideological
beliefs need to be inferred, hypothetically reconstructed, from actual discourse, for instance by comparison with repeated (contextually different)
discourses of other group members. Since we only have one (large) text
here, such comparisons can only be made within the book itself, as well as
with those texts or examples the author refers to and agrees with. Moreover,
in typical ideological treatises of this kind, the very formulation of the
'principies' involved may be close to the underlying ideologies because
D'Souza does not tel many concrete stories, but argues at a general, abstract
level. Moreover the overall, contextual purpose of the book is to attack what
he sees as a threatening ideology (cultural relativism) and to promote
another, which he does not narre explicitly, although he aligns himself with
what he calls'cultural conservatism'.
As may be expected from a book that deals with various political, social,
economic and cultural issues, also D'Souzá s book manifests several, related
ideologies, depending on lis respective identifications with different groups
or communities, as explained aboye: Western, white, middle-class, male,
heterosexual, professional, conservative elites. However, D'Souza focuses
on his main ideological and social enemies, namely, the cultural relativists
and African-Americans. Also class is a salient dimension, as is obvious from
his special wrath against the black 'underclass'. His frequent generalizations
show, however, that he takes the whole black community as a metonymic
(totum pro parte) representation of the black poor.
In sum, we may expect four types of ideology here, those of raceethnicity, class, culture and politics, and an overall 'meta-ideology' organizing these, namely, that of conservatism. It is this over-arching conservative
ideology that establishes coherente and numerous links between the beliefs
in the respective ideologies. For instance, where D'Souza defends sociopolitical, neo-liberal beliefs about limited state intervention, we may expect
racialized beliefs about African-American dependency on the state in
general, and about black welfare mothers in particular. And where his
cultural ideologies defend the uniqueness and hegemony of Western civilization, we may expect both the class and yace ideologies to feature beliefs
about the 'barbarism' of the underclass. The same cultural ideologies may be
connected to ideological beliefs about the 'bankruptcy' of relativist multiculturalism, whereas conservative—liberal individualism emphasizes the
importance of personal merit against group-based, collective affirmative
action. Similarly, the conservative ideology of law and order will be
'racialized' in this case in the evaluation of Ilack crimé . Many other such
The ideology and discourse of modern racism
285
cross-linkages between main ideologies and specific attitudes may be
reconstructed from this book.
As we shall see in more detall below, such an ideological complex will be
brought to bear in the central attitude that provides the basis and the title of
this book, namely, that contrary to what is maintained by blacks and their
liberal white supporters, the USA is not (or at least no longer) a racist
country. As suggested in the previous chapters, it is this denial of racism
which constitutes one of the core attitudes of modem elite racism. Disguised
by what is defined as a 'culture war' between liberal relativists and
conservative cultural supremacists, we thus discover the continuation of the
ongoing'rue war' that has characterized the'American dilemmá for
centuries. Indeed, the book's subtitle advocates a'multiracial' society, but
the contents of the book show that the supremacy of the dominant white
should not be challenged. 'Racional discrimination' is a 'natural' right
of this dominant ethnic group:
[141 The Greeks were ethnocentric, they showed a preference for their own. Such
tribalism they would have regarded as natural, and indeed we now know that it is
universal. In some situations an instinctive ethnocentrism is inevitable, as when
one's society is under extemal attacks and one must rally to its defence.
(p. 533)
We see here at work one of the most prominent devices of the ideological
legitimation of inequality, namely, that such a situation is 'natural' and
hence'universal'. At the same time, such a passage shows another device in
the representation of the others, namely, how outgroups are constructed as
enemies against whose 'external attacks' we must 'naturally' defend ourselves. Thus, racism is not only made respectable, while natural, but also a
patriotic duty of whites in the 'culture war' and the 'civilizational crisis'
(p. 535).
After this brief overall characterization of the various ideologies involved,
let us now examine some of their contents and structures.
Conservatism
It was argued, aboye, that 'conservatism' it not so much a (group) ideology,
but rather an overarching, meta-ideology that organizes other ideologies. For
instance, applied to neo-liberal ideologies in the realm of the political
economy, conservative ideologies typically advocate a limited role of the
state (or government) in the market. Similarly, when applied to cultural
ideologies, conservative meta-principles may take two complimentary variants: limited state intervention in some cultural domains (education, media,
religion), or active state intervention, for example through tough legislation,
in the domains that are seen to threaten the moral order (family values,
sexuality, multiculturalism). And finally, when applied to racial or ethnic
ideologies, conservatism will similarly allow (condone or not strictly police)
various forms of discrimination, as the right of each person or ethnic group
to 'prefer one's own'.
286
Discourse
Values As all ideologies, also conservative meta-ideologies are based on a
selection and combination of values drawn from a cultural commonground.
D'Souza for instance positively refers to the following values (of which the
ideological, attitudinal and discursive constructions will be examined
below):
• freedom
• personal merit
• discipline
• prudence
• moderation
• responsibility
• self-restraint
• hard work
• authority
• order
• decency
• elitism
• non-permissiveness.
Such an ideological selection of rather general cultural values usually also
involves a set of counter-values when the ideology is brought to bear in an
ideological struggle with ideological opponents. Thus, these values are
selected and emphasized especially against (certain variants of) those of
egalitarian, progressive liberalism: equality, social responsibility, social
support, moral freedom, cultural relativism, freedom from oppression,
representativeness, anti-authoritarianism, permissiveness, creativity, selfcritique, progress, democracy, and so on.
Given these values and their counterparts, sorne of the conservative
ideological beliefs defended by D'Souza in his book are the following.
1 The social and civilizational status quo is being threatened.
2 The state should not interfere where ft does not belong.
3 Social programmes to help the poor are counterproductive.
4 People should be judged individually by their own achievements.
5 Inequality has individual not social causes.
6 People have duties, and not only rights.
7 A cohereñt society does not allow multiple cultures or worldviews.
8 There are natural inequalities between (groups of) people.
9 Society must be characterized by law and order.
10 All individuals should take initiative and pursue excellence.
11 Children shall be born in wedlock.
12 All people must work.
These ideological principles are not always directly formulated in the End
of Racism, but especially appear in the negative evaluation of the ideologies
and attitudes of D'Souzá s enemies, for example in favour of state intervention in the ghetto, welfare, affirmative action, social responsibility of
The ideology and discourse of modern racism
287
business companies, social disadvantage, the legitimacy of single mothers or
other family structures, decent jobs, equal group representation, equal
outcomes, and so on.
As suggested before, these conservative values and ideological beliefs
will appear to be manifested in more specific group ideologies and attitudes.
Indeed, some of the ideological beliefs mentioned aboye might even be
omitted because they are domain- or group-specific general beliefs. Thus,
the freedom from state intervention in fact implies that the state should also
not be (very) active in the social domains, for example with social programmes for the poor or the elderly. Similarly, the opposition to 'illegitimacy' of children or to unmarried mothers, is of cocarse a further
specification of overall conservative beliefs about family values.
Ethnocentrism/modern racism
Although conservatism is the overarching ideological framework that organizes the social and cultural beliefs in The End ofRacism, ethnocentric
modem racism is its specific ideological core. This conclusion may be rather
ironical given the title of D'Souzá s book, but within the frarnework of our
elite theory of racism, such denials are paramount in all forms of modem
racism. Hence D'Souzá s rage against anti-racists, his systematic mitigations
of the continued relevance of'racé in the USA, and his alleged 'ignorance'
of widespread discrimination against of African-Americans in virtually ah
social dornains. For the sarne ideological reasons he attacks civil rights
'activista', those who plead for (or see no alternative for) affirmative action,
and those he sees as using racism as an excuse for own failure and
'civilizational breakdowñ .
As group ideologies, ethnocentrism and modem racism feature the following basic beliefs about the own group, namely (white) Westerners, and its
relations to other groups. Most of these ideological principles are based on
the core value of (cultural if not natural) inequality between groups.
1 Our Western culture is superior.
2 Ethnocentrism is natural and sometimes inevitable.
3 Discrimination may be rational.
4 The USA is not and should not be a multicultural society.
5 Cultural assimilation of culturally deviant groups is necessary.
6 We are tolerant.
7 The USA is not a racist society. / We are not racists.
Related to these ideological self-representations is the, polarized, negative
representation of the others: first the liberal cultural relativists, for example
in tercos of the following beliefs.
1 They think that all cultures are equally valuable.
2 They advocate multiculturalism.
3 They criticize Western civilization.
4 They accuse us of colonialism and racism.
288
Discourse
5 They want proportional representation of ethnic minorities.
The second main enemy, the social opponents, are blacks, AfricanAmericans, and more generally all non-Westemers in the book. They are
variously described on the basis of the following ideological beliefs.
1 They are primitive, uncivilized, barbarians.
2 African-American pathologies are cultural.
3 They are culturally deviant.
4 They break the law.
5 They tend to be criminal.
6 Their culture(s) are stagnant.
7 They depend on the state.
8 They take no initiative.
9 They are promiscuous.
10 They are not striving for excellence.
11 They use racism as an excuse for own failure.
In other words, and as we have seen in Chapter 6, negative otherpresentations deriving from ethnocentric and racist ideologies are often
articulated around the attribution of violations of our basic values and
ideological principies. Thus, where we are tolerant, anti-racism is intolerant;
where we value personal merit and discipline, they lack such values; where
we are decent they are promiscuous, where we work hard, they are too lazy
to work, and so on.
Ideological structures
One of the theoretical issues dealt with earlier in this book is that of the
structure of ideologies (Chapter 5). On the basis of repeated general
propositions in D'Souza's book, a number of beliefs were selectál that are
general enough to be included in the conservative meta-ideology and the
ideologies of cultural racism or ethnocentrism. It was, however, argued that
ideologies probably have some kind of internal organization, for instance a
schematic structure of fixed categories. Such a schema would be relevant
each time people need to acquire or change an ideology, for instance when
they become new members of a social group. Searching for a format for
such a schema, I assumed that given the close link between group ideology
and the self-representation of the group, a group schema modelled on the
fundamental societal co-ordinates of the group would be a good candidate.
The question now is whether the ideological propositions inferred from de
D'Souza's book can be validly assigned to such a schema.
Thus, if we have to design a framework for the ideologies of racism and
ethnocentrism, we may propone the following (simplified) structure:
• Membership Criteria — only members of our own culture, ethnic group,
race' or nation;
• Activities discriminate others;
• Goals — exclusion, segregation or assimilation of others;
—
The ideology and discourse of modem racism
289
• Values — natural inequality, cultural homogeneity;
• Societal Position: relation to other groups — we (our culture) are (is)
superior to the others;
• Resources — Western civilization, (political and economic) power,
whiteness.
Obviously, sine group self-schemata are usually (though not always)
positive, and 'racism' is culturally and socially sanctioned, at least officially,
most people who share this schema will not describe themselves as 'racists',
but for instance as nationalists Recall that the group schema and its
categories will feature those fundamental group beliefs that define the
identity as well as the basic interests of the group. When these interests are
under threat, they will most energetically be defended, or when lost they will
be reclaimed.
This is also the case for D'Souza' s book. Thus the membership criteria
category defines who does or may belong to Us, and hence the others are
defined by racists or ethnocentrists as foreigners, aliens, immigrants, outsiders, and so on. The activities of the members should be geared towards
the realization of the essential group goal, which is basically to keep others
out or down, or if that is impossible to fully assimilate them (in this case
culturally). These aims are the basis of the negative evaluations in the
attitude of multiculturalism, as we shall see below.
The basic value of ethnocentrism and racism is to emphasize'natural'
inequality between groups, against the egalitarians and the relativista. It is
not surprising that such a value only serves the interests of those who are
dominant, and therefore, in the societal position category, we find the
fundamental definition of Our position, namely, that We are superior to
Them (i.e. Our civilization, culture, knowledge, etc., is better than
Theirs).
Since dominant group position and reproduction need resources, the
crucial resource in a racist ideology is the symbolic power of being part of
(Western) civilization and of being white, that is, the very criteria of their
membership of their own group. Given the fundamental nature of resources
for group power and reproduction, these are the ideological interests that
will be defended most forcefully. This is indeed the case in D'Souzá s book,
wherein the repeatedly expressed concem is that (Western) civilization is
breaking down, that other cultures may get the upper hand, and that Our
(Western, white, male, middle-class, etc.) group and its interests may lose
power.
The societal position category in the ideological schema typically features
a relation to other groups, in this case obviously the group(s) that are the
very target of racist or ethnocentric groups, namely, foreigners, immigrants,
aliens, minorities, and so on, especially those of another culture and/or
appearance ('race ). Given the relationship of superiority involved here, the
other-group schema associated to this self-schema typically will feature
those categories and beliefs that are opposed to those for our own group.
290
Discourse
Their membership category (as defined by Us) is, say, 'being black', or as
D'Souza's insists, 'having a coherent black culture'. It is also here that the
'essential' evaluation of the others is being represented, namely, as being
primitive, uncivilized, barbarians, lacking initiative, being promiscuous, and
so forth.
Their (negative) activities may be ideologically summarized as 'They
violate all our norms' (are criminals, push drugs, get illegitimate children,
don' t want to work, accuse us of racism, etc.). Their goal is represented, for
example, as equal rights, multiculturalism and an equal economic share.
Their values are all those opposed to Ours: egalitarianism, relativism,
permissiveness, dependency on state, disrespectfulness of order and authority, indecency, and so on. Their position is represented on the one hand as
(culturally) inferior, and on the other hand as a threat to our culture,
civilization and other resources; moreover, they accuse us of racism and
intolerance. Since the other group is hardly powerful, few resources will be
attributed to them, and the point is precisely to make sure that they will not
get access to our resources, or their resources (such as their own culture) will
be negatively valued, as is the case, as we shall see, for the attitude about
Afrocentrism.
These basic ideological group schemas for Us and Them will then be
further detailed for specific social domains in a number of more detailed
attitudes about specific groups, for example about African-Americans, or
about Us (whites, etc.) in the USA, and for specific issues, such as racism,
multiculturalism or affirmative action, as I shall spell out below.
Note, finaily, that I did not attempt to schematize the list of conservative
basic beliefs, since conservatism is not a specific group ideology, but rather
a meta-ideology that organizes some basic principies of other group ideologies. The typical conservative beliefs (about state intervention, individualism, law and order, family structure, etc.) are in fact all specifications of
fundamental conservative values. Thus freedom is defined as freedom from
state intervention, and personal merit is inconsistent with social welfare,
decency prohibits illegitimate children, and so on. If we would have to
define conservatives as a 'group' we might say that it is constituted precisely
by the category of its values (against progressives). That is, the identity,
actions, goals, position and resources of conservatives all focus on the
realization of those values. It is in this way that the meta-ideology of
conservatism constrains other (group) ideologies, such as those of racists, or
professors, or business people, for whom the conservative value system will
have different applications depending on the interests and specific group
goals of these groups.
Altitudes
Theoretically, ideologies control and organize more specific attitudes. Thus,
whereas basic ethnocentric and racist ideologies represent the overall
The ideology and discourse of modern racism
291
properties of Us (Westerners, whites) and Them (non-Westerners, blacks),
attitudes feature more specific social beliefs, such as prejudices, about
specific outgroups. Thus, African-Americans are further represented as
follows.
1 They are the cause of the breakdown of civilization.
2 They have one coherent (black) culture.
3 (Poor) blacks have scandalous pathologies:
• excessive reliance on government;
• conspirational paranoia about racism;
• resistance to academic achievement;
• celebration of the criminal;
• normalization of illegitimacy;
• single-parent families.
4 Their pathologies are due to African-American culture.
5 Their culture is functionally inadequate.
6 They are themselves racist:
• they have ideology of black supremacy.
7 They are violent and criminal.
8 They abuse drugs.
9 They have an expensive lifestyle (they are Ilashy').
10 They may have lower intelligence.
11 They have fewer skilis.
12 They have no mores.
13 They celebrate or condone broken families
14 They do not adapt to the dominant (Our) culture.
15 They do not take responsibility.
16 They have paranoia about racism.
17 Their middle class has an unfounded black rage.
18 They are weak in developing businesses.
19 They repudiate standard English.
20 They celebrate the Sad Nigger'.
21 They dress in conspicuous clothes.
22 They use obscene language.
23 They do not want to work.
24 They are not puntual.
25 They do not respect matrimony.
26 They cause the bastardization of America.
27 Their intellectuals refuse to criticize underclass pathologies.
These beliefs may be further organized in a more structured schema of
which, however, the overall principie is again clear: the others (here the
blacks) are represented as our negative mirror image — literaily as our dark
side. Whatever values and principies We share, They don't have them.
The core concepts organizing these beliefs are d jerence, deviation and
threat, applied in all social domains, for example those of culture in general,
habits, language, dress, work ethic, family values, character, tolerance,
292
Discourse
modesty, industriousness, individual merit and achievement, and so on. That
is, their cultural mores are not only different from ours, but they also deviate
from our norms and laws, and ultimately, their cultural deviance as well as
their aggression, crime and other behaviour are a threat to Us and the whole
nation, including themselves. Note that within the attitudinal representation
of African-Americans, we also encounter some specific beliefs about black
subgroups, such as black intellectuals, wornen or the 'underclass'.
Often, however, the text is not that specific, so that many negative
attributes ascribed to a relatively small group of young men in the ghettos
are in fact generalized towards the whole group. It is this (over)generalization that is one of the hallmarks of racism: they are all alike. Although
D'Souza recalls (without much conviction) that is not in their genes', and
that he therefore cannot be called a racist, the distinction between AfricanAmerican 'culturé and'racé is very subtle in his argument, and often nonexistent. Indeed, most blacks would see his very negative and aggressive
stereotyping as little more than a forro of racist derogation hiding behind a
thin veil of cultural critique.
In his rejection of racism as the cause of the deplorable social condition of
the African-American community, D'Souza has no other option than to
blame the victims themselves (a strategy he energetically denies and even
attacks as one of the criticized forms of anti-racism). That is, he focuses on
Ilack pathologies' and sees these as a 'civilizational breakdown', as
discussed aboye. Hence the blacks, and no other group or organization, are
the cause of the 'catastrophe' that is threatening 'Us' in the USA.
More sober analysts of the socio-political situation in the USA (and
elsewhere in the world) would probably wonder why D'Souzá s rhetoric
focuses on just those 'pathologies' and why these should constitute something as dramatic as a 'civilizational breakdown' and a 'threat' to the whole
nation. Since when is welfare, when no jobs are available, a pathological
forro of 'parasitic reliancé ? If so, most of the Western European welfare
systems would not be an object of envy. And what about single-parent
families? These are increasingly normal in many parts of the world,
especially in highly developed nations, such as those in Scandinavia, where
up to around 40 per cent of mothers are not married. What we have here,
obviously, is a socio-cultural difference, and hardly a pathology, and even
less something as apocalyptic as the 'bastardization of America' as D'Souza
so delicately describes black families And how would D'Souzá s black
conservative friends who are prominent professors (as well as all other
blacks with an academic degree) interpret his conclusion that AfricanAmericans are 'hostilé to achievement? Surely, there are other, more
fundamental, social and economic problems in the USA, such as the poverty
of many millions of families and children.
What is important for my analysis, however, is not so much a critical
challenge of D'Souzá s work (many others have done that already) but a
demonstration of how values, ideologies and altitudes influence the definition and evaluation of the social situation. Where many see poverty, racism,
The ideology and discourse of modem racism
293
marginalization and many other social ills in the USA, D'Souza's ideology
has blinded him to such realities. On the contrary, in a grand movement of
reversal he blames the victims of this situation. Even a well-founded
analysis of US society, not only by blacks, is thus claimed to be pathological. Hence, we see how different ideologies may lead to opposed
assessments of the'facts'.
Attitudes about racism
D'Souza's ideologies also control attitudes other than those about AfricanAmericans: for instance, as we have seen, about racism. Again, both his
knowledge and opinions about racism appear to be heavily biased by his
underlying ideology of ethnocentrism and modem racism. First, however, it
is crucial that his opinions about cace and racism be safely protected against
any accusation of racism. He does this, as is usual in much other elite
discourse, also among several social scientists, by limiting the definition of
racism to a 'belief in intrinsic, biologically based superiority'. Since only
small groups of white supremacists share this belief, his beliefs and those of
most other modem racists are safeguarded against any accusation of racism.
Racism defined as he does, is indeed a marginal problem in the USA or
anywhere else. The problem is that the system of ethnic/racial inequality in
the USA (and other countries dominated by Europeans) is much more
complex than that, and not limited at all to beliefs about biologically based
superiority. Rather, especially when associated with appearance, all feelings
of group-based superiority, also those of culture, and the many everyday
forms of discrimination based on them, are forms of contemporary racism.
The same is true for most other beliefs about racism D'Souza expresses in
his book. That is, they are geared towards protecting himself and dominant
white culture and civilization from the uncivilized taint of racism. Thus, a
whole chapter is dedicated to a historical treatise about racism (and slavery)
as existing in many other countries and civifizations, and concludes that
white Europeans were not the only 'guilty' ones. Such a chapter should also
be interpreted as a tactical move to at least share the blame of racism. And
when the title and much of the content of bis book emphasizes (correctly)
that racism is not universal, but has a specific beginning and end, he (falsely)
concludes that (therefore?) racism in the USA has ended (on a par with
ideology and history which other influential conservative authors before him
declared to have 'ended). Once established (without proof, and disregarding
libraries full of evidence to the contrary) that racism has ended, the real aim
of the argument becomes clear: if there is no significant racism in the USA
anymore, the blacks can safely be blamed themselves for their 'pathologies',
and We (whites) are again in the clear. And even more forcefully, those may
be accused of bias or lies (or worse, reaping'profits) who claim that racism
is alive and kicking today in the USA as long as we do not limit its
definition to marginal phenomena such as beliefs in 'biological superiority
of the white racé . Thus, where D' Souza claims that 'accusations of racism
294
Discourse
are a rationalization of black failure', others may reverse the claim and state
that D'Souza's denial of racism is a rationalization of continued white
failure to come to terms with blacks in US society. No wonder that in
D'Souza's hierarchy of ideological values, the real problem is not racism,
but anti-racism — defined as 'intellectual and moral coercioñ .
In D'Souza' s attitude to racism, even when defined in his way, racism is
a legitimate opinion (p. 538), which may be criticized, but which is no
crime, despite many international laws, United Nations charters against
racism, and the Universal Declaration of Human Rights. We see that at this
point the conservative values of law and order clash with the principies of
his racist ideology — enforcing the many laws against discrimination does
not exactly have priority in this attitude. On the contrary, discrimination, as
D'Souza argues, may well be rational and legitimate in some situations.
Ethnocentric supremacy, and neo-liberal freedom to discriminate (e.g. in
business) are ideologically superior to the liberal principie of not violating
the rights of others.
Wherever discrimination and racism cannot be bluntly denied it is
mitigated, their current relevance and seriousness played down, or even
legitimated in specific situations. Systematic everyday discrimination in the
USA is thus euphemistically reduced to such improprieties as 'slights of
taxidrivers who pass by African Americans' (p. 525), a forro of discrimination that is fully legitimate for D'Souza, because it is 'rational'.
Similarly, although structurally very similar and socially equally destructive, US segregation is deemed to be totally incomparable with apartheid, a
familiar move of mitigating denial. And when D'Souza claims that 'we do
not know how much racism exists in the USA', such a well-known move of
apparent ignorance ('nobody knows how to measure it) is curiously
inconsistent with his own repeated claim that racism has declined in the
USA. But should some racism still exist, it is especially due to the behaviour
of the black underclass, which violates all social and cultural codes of US
society — another reversal by blaming the victim.
There are several points where D'Souza's beliefs about racism coincide
with those of critical scholars who have studied racism. Thus, as we have
seen, racism is certainly not universal, but a scientific invention of
eighteenth- and nineteenth-century Europeans, for example, used to explain
observations of the 'primitivism' of other cultures. D'Souza does not
mention here that it was also invented to legitimate slavery, genocide, land
grab, colonization and many other highlights of Western 'civilizatioñ .
Indeed, racism is not an irrational antipathy of stupid, uneducated people,
but had a scientific basis (and such science should never be called'pseudosciencé warns D'Souza). True, as is exemplary for his own book, racism,
ethnocentrism and many other forms of inequality have always been
preformulated and legitimated, in more or less respectable academic terms,
by the elites. Por D'Souza, however, the argument has other implications: if
discrimination is not irrational, but rational, this means that (white) people
may have good, even respectable, reasons to discriminate against blacks.
The ideology and discourse of modern racism
295
Thus, D'Souza defends the attitude that prejudice, discrirnination and
ethnocentrism may be natural, rational, expedient (good for business) and
hence justified.
Given his attitude to racism, we should not be surprised by D'Souza's
attitude to colonialism, another invention of Western 'civilizatioñ . Denial,
mitigation, legitimation and simply ignoring the historical facts are only
some of the strategies employed to protect the ideology of Western civilizational supremacy against overly critical examination. Thus, explorations
were not 'carried out with hostile intentions', D'Souza claims, and should
not be seen as rapacious land grabs, theft of resources, or (sometimes)
genocide, but as Europé s contribution to 'world transformatioñ, as a sigas
of progress, and as intellectual enterprises. What is clear from such attitudinal beliefs is that ideologies have a very powerful control over the very
selection, focus, representation and construction of historical'facts'. And
where ideologies, such as that of Western civilizational supremacy, might be
inconsistent with these facts, they may be insulated against these facts by an
entirely different version of reality.
Affirmative action
Little speculation is necessary to predict D'Souza's attitudes about affirmative action (AA), given his denial of racism and his conservative values and
ideologies of personal merit, discipline, hard work and rejection of any
government intervention. Whereas, on many other accounts, D'Souza rejects
egalitarian values, social policy should, according to him be 'colocar blind'.
He insists that this principle of Martin Luther King should be respected, but
he especially does so to demonstrate that contemporary black intellectuals
violate King's legacy: a well-known tactic of dividing the enerny.
Whereas elsewhere in his book he makes a case for the legitimacy of
rational' discrimination, affirmative action is strictly rejected because it is
defined as discrimination — of whites that is. Following his own criteria that
allow discrimination, one might ask whether affirmative action is an
Irrational antipathy', rather than a rational policy to end inequality and
many remaining disparities in hiring, promotion and work conditions of
minorities in general, and blacks in particular.
That AA would corrupt US firms, as another of D'Souza's attitudinal
beliefs suggests, is another definition of the situation biased by the fundamental ideological belief that social and ethnic inequality should not be
taken very seriously. It certainly does not explain why many big companies,
when free to decide whether to apply AA, choose to do so.
The most familiar attitudinal belief about AA is that it would lower
standards', which presupposes that minorities (and especially blacks) are
generally less qualified. D'Souza extensively cites all statistics to prove just
that. Since he has rejected a 'racial' (biological) explanation of such lower
qualification, he is free to play the 'culture card', and hence accuse blacks of
lacking a culture of achievement. Of course, other social explanations (bad
296
Discourse
schools) are hardly highlighted, nor may we expect the conclusion that if
blacks (as a group) perform less well, not they, the victims, but the schools
should be blamed.
Finally, in this sequence of accusations, reproaches and blaming the
victim, D'Souzá s attitude on AA, as well as on civil rights more generally,
is that this system means Big Bucks for the 'civil rights industry', and
especially also for black intellectuals and (other) 'activista; how much, he
does not tel us, one of the many claims about which suddenly his footnotes
are scarce, nor how much white civil servants profit of the system. That
attitudes are not always consistent among each other, shows here, because
where the 'Civil Rights Industry' (and their black employees) are accused of
gobbling up mega-dollars, blacks are elsewhere accused of not taking
enough corporate or financial initiatives, This 'pathology' D'Souza traces
back to black attitudes during slavery: 'a series of measures to avoid,
postpone and minimize work' (p. 97). In more traditional parlance such an
attitude was routinely expressed as They are lazy.' The point, thus, is not to
try to establish a balanced picture of the social situation of African
Americans and yace relations in the USA, but to fmd any argument to
derogate blacks.
Multiculturalism
The ideology of cultural conservatism is not very friendly towards multiculturalism. As D'Souza also has shown in his earlier work,' in which he
ridicules educational, curricular and scientific diversity, his ideology of
Western cultural supremacy is inconsistent with the cultural relativism of the
'Boasians', and with that of most social scientists in the world, for that
matter.
The specific attitudinal cluster organizing his beliefs about multiculturalism is organized by a number of familiar dimensions, such as the conceptual
triple, encountered before for the representation of African-Americans:
difference, deviance, threat. Multiculturalists are different from us, deviate
from our cultural and educational norms and are even a threat to our Western
civilization.
To make the case for the 'threat', various devices of hyperbole are of
course necessary4 as has been the case more generally in the debate about
multiculturalism. In such an attitudinal framework it is not consistent, for
instance, to consider alternative versions of reality, for instance the fact that
multicultural education in US schools, colleges and universities is, as yet,
marginal with respect to that of the teaching and research about dominant
Western culture from Aristotle, to Shakespeare and Einstein.
Another ploy to emphasize the deviance of multiculturalism is to associate
it with other evil cultural developments, as seen by the cultural conservatives: Marxism, deconstructionism, and Third World scholarship, none
of which is exactly a dominant force in US academia. However, appeals to
anti-communist (Le. un-American), ideologies, ethnocentric doubts about the
The ideology and discourse of modern racism
297
excellence of Third World scholars, and anti-intellectual ridicule of (also
foreign, while French) deconstructionism, are of course consistent with both
conservative and ethnocentric ideologies.
That multiculturalism would 'result in imbalance and distortion', as
another belief of this attitude states, is a final strategy in the negative
representation of curricula that emphasize the need for educational diversity
for an increasingly ethnically varied population. Of course, the imbalance of
restricting education largely to Western authors and scientists, is not further
considered, nor has D'Souza nor other representatives of cultural conservatism written alarming books about this form of scholarly distortion,
which has dominated US (and other Western) education until today.
Afrocentrism
The combined ideologies of cultural conservatism, ethnocentrism and modem racism are brought to bear in the construction of an extremely negative
attitude about Afrocentrism. Ridicule, over-generalization and hyperbole are
also the major strategic moves here. Afrocentrism is thus represented as a
dangerous philosophy. As usual in the representation of blacks, the views of
a radical minority are first generalized and exaggerated by selective quotes,
and then derogated. Altemative representations of Afrocentrism, as a correction to dominant Eurocentric ideologies, and as a means to enhance group
identification and pride among blacks, would imply a relativist position that
is of course inconceivable for D'Souza.
Where arguments about scholarship, the arts or other elements of culture
do not suffice, there is always the option to ridicule and derogate the
appearance or behaviour of blacks who indulge in Afrocentric beliefs,
following the familiar constraints of modem racism:
[15] ... the hardened gleam in many Afrocentric eyes ... virtually cultic pattem
of lockstep behavior: eve one dresses alike, and when the leader laughs,
everyone laughs.... (p. 381)
Depending on oné s ideology and social actitudes, such a description
would of course fit many outgroups, ranging from the military to the
denizens of Wall Street. That is, there is no aim to correctly describe the
others, but to construct a negative stereotype, according to which others are
typically'all aliké and lack humanity, individuality and autonomy.
The IQ debate
Finally, D'Souza engages in a lengthy discussion of che IQ debate, spawned
by the controversial book by Hernnstein and Murray, The Bell Curve. His
position is very ambiguous here. He feels ideologically related to there
authors because they also question 'the foundation of twentieth century
liberalism: the denial of natural differences and the premise of the inherent
equality of groups' (p. 434). Indeed, he asks-
298
Discourse
[16]Why should groups with different skin color, head shape, and other visible
characteristics prove identical in reasoning ability or to construct an advanced
civilization? (p. 440)
He thus goes a long way towards agreeing with many of the racist
presuppositions of Hemnstein and Murray. He extensively cites all scientific
evidence that claims to show genetic black inferiority, as he also does when
discussing the presumed lack of academic achievements of blacks. He
thereby ignores the vast literature that shows that intelligence is largely
contextual and socio-economical, and may even change dramatically within
the same group within one generation. This shows that, typically for
ideologically based persuasion, evidence is selectively focused on and
presented in accordance with oné s group attitudes: only those data that
confirm the negative characteristics of the others will be given due attention.
The whole argument of biologically based racial differences of intelligence (and culture), of course presupposes the viability of the very notion of
race', which he claims most scientists accept:
[17]Most anthropologists and biologists agree on the existence of three broad
racial groups: the Caucasoid, the Negroid and the Mongoloid. (p. 449)
Again, he is virtually silent about (or simply rejects) all the scholarly
literature that concludes that, despite obvious and undeniable differences of
appearances between people in the world, a classification of people into
races' on the basis of such (superficial) differences of appearance only
makes sense in common sense. It is the same common sense, rather than
scholarly evidence that makes D'Souza smugly use the following argumentum ad absurdum:
[18]If the concept of race is entirely fictional, shouldn't all civil rights laws which
rely on racial classification be struck down by the Supreme Court as meaningless
and unconstitutionally vague? (p. 447)
One of the many problems with this argument is that he disregards the
difference between a biological classification and a socio-political or legal
one. 'Race' is a social, commonsense construct, and racism is based on such
a commonsense classification. Legal measures to counter racism of course
recognize the existence of a social category of on which racism is
based, but do not presuppose the existence of biological classifications of
people into races.
Again, the strategic aim of D'Souza's argument is not so much to prove or
disprove the existence of biological races, but rather to provoke supporters
of civil rights, and hence his ideological enemies, into accepting biological
races through the back door of the social and legal classifications of
Moreover, the argument is inconsistent with his critique of the 'one drop of
blood rule' that (socially) defines people in the USA as 'black' if they have
one drop of 'black' blood. If, indeed, most blacks in the USA do have
'mixed' ancestry, then the very point of their biological classification as
black (and hence their racial inferiority on IQ tests) makes litde sense. Thus,
The ideology and discourse of modem racism
299
what African-Americans do have in common, though, is their social position, namely, as being self- and other-defined as being black.
Ultimately, however, D'Souza rejects (without'much argument) the biological account of 'racial' differences and the IQ gap between blacks and
whites, because that would be inconsistent with his ideologies of cultural
conservatism and ethnocentrism. After all, if the 'pathologies' of the black
community were largely caused by their genetic predisposition, they could
hardly be blamed for them. A cultural explanation, by which deviant black
culture is seen as the source of all problems, is much more persuasive in an
argument that sets out to emphasize Western, white civilizational supremacy. Such an argument also rules out, as we have seen, any socio-economic
explanation of African-American'failuré :
f 19] My conclusion is that it is an illusion to 'believe that racial differences
etween blacks and whites are largely a phenomenon of socioeconomic class and
that such differences will disappear with the current menu of preschool and
public-school government interventions. (p. 457)
L20] Contrary to the assumption of cultural relativism, the problem, it seems,
is not test bias but the functional inadequacy of African-American culture.
(p. 461)
These discursive manifestations of underlying attitudes show again how
beliefs are strategically shaped in accordance with prevalent ideologies.
According to ethnocentric and modem racist ideologies, blacks need to be
represented as inferior to whites. The cultural ideology then provides the
explanation of such inferiority in terms of the 'functional inadequacy of
African-American culture', which again is the belief that sustains the
vehement attack against African-Americans. Biological explanations of
black inferiority would invalidate such an argument, although D'Souza
seems quite impressed by the biological evidence that might explain the IQ
gap as well as the cultural inferiority of blacks. However, if D'Souza were to
accept that blacks are genetically unable to compete with whites (or Asians),
one solution would again be affirmative action and remedial schooling, and
hence (more) government intervention, which is of course off limits for the
conservative ideologue.
Models
We already briefly indicated, aboye, that ideologies and the social attitudes
they control not only appear directly in discourse, as general statements, but
also affect mental models, that is, personal interpretations and opinions
about concrete events. D'Souzá s book has few stories of such personal
experiences: the 'definition of the situation' he presents is generally quite
abstract. However, when he does tell about such an experience, we do see
how underlying ideologies also control his mental models. Here are small
fragments of one of the stories that express such a personal model, namely,
his experience of the celebration, on 28 August 1993, of the thirtieth
anniversary of Martin Luther King' s march and '1 have a dream' speech:
300
Discourse
[21] ... One by one the leading civil rights spokespersons took the podium,
gravely invoked the memory of Martin Luther King, Jr. and demanded that
Americans do more to vanquish the forces of white racism so that blacks could
achieve what one speaker termed 'meaningful eguality'. .
But I did not hear anyone invoke King's principie of a race neutral society in
which laws and policies are indifferent to colour. The reason for this reluctance
was implicitly expressed by black activist Benjamin Chavis's rallying cry. 'We
don't just want equal rights,' he said. 'We want our fair share of the economy.'
Other speakers decried what they termed 'institutional racism', although they
were not specific about this terco. The rhetoric suggested the existence of a new
civil rights agenda, in important respects different from the one which Martin
Luther King, Jr. championed...
Certainly the style and tone of the 1993 assembly differed in two important
respects from that of King's march three decades earlier. First, many of the
audience seemed middle-class, and diere were conspicuous signs of prosperity. A
number of speakers arrived in chauffeured cars. I overheard talk of appointments
and schedules. 'I have to be at the coalition meeting at six.' I hope that they hold
my dinner reservation.' ... Some activists engaged in minor turf warfare, sparring
over whether they had been booked at the Willard or the Madison hotel, over who
spoke first at the podium, over who sat where on the dais, and so on. One black
professor who felt neglected erupted,'This event replicates the structures of
oppression in American society.' Despite this distress, it was gratifying to see
indications that the lives of many blacks in the United States have nnproved
dramatically. People whose condition is economically and socially desperate do
not fret over speaker schedules and hotel bookings. (pp. 201 _2)
As this passage shows, his personal model of the event closely follows his
general attitudes about the condition of black America: a successful black
middle class which do not cace about the black underclass, black people
being 'conspicuous' with their wealth, and attributing ah problems to
racism, whereas the 'real' problem is the violence of the ghetto. Strategically
aligning himself with Martin Luther King (a move of positive selfpresentation), he sees the manifestation as a contradiction to what King and
D'Souza favour: a'race neutral society'. The description and the ironic and
derogatory style of this story obviously define the event in tercos that are
consistent with this attitude about the black community Prominent in that
attitude is the rejection of racism as the main problem of black America, and
an emphasis on violence and other 'pathologies' of the inner cides and of 'a
second black America'. That is, not our failure (racism) but their failure
(pathologies), are then explained in terms of an overall 'black culture'
(p. 204). Such an overall classification ignores class division so that the
whole black community can be blamed, as he also does in the passage just
quoted. In other words, the ideologies of modem racism and cultural
conservatism combined produce a mind-set that has such biased models as a
result: D'Souza can only 'see' the events the way he describes them.
Discourse
Social representations and personal models control the style and content of
text and talk. Let us therefore finally examine how D'Souzá s ideologies and
The ideology and discourse of modern racism
301
attitudes, as well as his personal views, impinge on the discourse of his
book. Space limitations do not allow me to provide a detailed discourse
analysis of a book of 724 pages, however, so I must limit myself to brief
comments on some significant passages. Since ideologies about groups and
cultures are involved, I shall focus on the well-known ideological square of
positive self-presentation and negative other-presentation. Also, the analysis
will be relatively informal so as to enhance its readability. Within the
broader framework of a critical discourse analysis, I shall also occasionally
formulate critical comments on D'Souza's book, but my aim is to illustrate
the relations between ideologies, actitudes, models and actual discourse,
rather than to denounce D'Souza's book or the ideologies that he represents.
Aboye, I have already given some text examples from D'Souza's book,
and also briefly made some analytical remarks about them. Typical for a
rhetorical book like this, which intends to contribute to an ongoing ideological debate, and which aims to sharply criticize the black community, are the
various strategies that represent Us and Them. Thus, Our group as well as
those with whom D'Souza identifies himself, namely, the West, Western
civilization, Europe, white people, conservatives, and so on, are consistently
described in positive tercos, whereas any negative characteristics will be
ignored or mitigated, whereas the others and especially African-Americans
are consistently described in negative tercos.
Derogating African-Americans
Thus black 'pathologies' are described in stark contrast to Us, and with the
usual forros of hyperbole. Lexical choice, comparisons, metaphors and any
other device that may be used to paint a negative picture of blacks will be
used. Here are some examples, which I cite at length in order to get a good
impression of D'Souza's discursive style (some repeating earlier quotes):
[22]... the hardened gleam in many Afrocentric eyes ... virtually cultic pattern
of lockstep behavior: everyone dresses alike, and when the leader laughs,
everyone laughs.... (p. 381)
[23]... black racism is more explicitly menacing. (p. 421)
[24] Louis Farrakhan reportedly uses the profits to subsidize a lavish lifestyle
which includes expensive silk suits and stretch limousine. (p. 426)
[25]The last few decades have witnessed nothing less than a breakdown of
civilization within the African-American community. The breakdown is characterized by extremely high rates of criminal activity, by the normalization of
illegttimacy, by the predo minan ce of single-parent families, by high levels of
addiction to alcohol and drugs, by a parasitic rehance on government provision,
by a hostility to academic achievement, and by a scarcity of independent
enterprises. (p. 477)
[26]The conspicuous pathologies of blacks are the product of catastrophic
cultural change that poses a threat both to the African-American community and
to society as a whole. (p. 478)
302
Discourse
[27]Of course no one is to blame for being a victim. But if as a reaction to being
victimized, a group develops dysfunctional or destructive pattems of behavior
which perpetuare a vicious cycle of poverty, dependency, and violence, then
continuing to inveigh against the oppressor cannot offer the victim much relie£
(p.482)
[28]Yet black culture also has a vicious, self-defeating, and repellent underside
Wat it is no longer possible to ignore or euphemize. As more and more blacks
seem to realize, no good is achieved by dressing these pathologies in sociological
cara, complete with the familiar vocabulary of disadvantage and holding society
to account. Society must do its part, and blacks must do theirs. But first, the
magnitude of the civilizational crisis facing the black community must be
recognized. This crisis points to deficiencies not of biology but of culture; yet they
are deficiencies and they should be corrected. (p. 486)
[29]For them [middle class blacks], apparently, antiracist militancy is carried to
the point of virtual mental instability. It is hard to imagine whites feeling secure
working with such persons: surely such inflamed ethnic insensitivitiel are now
what companies have in mind when they extol the diversity of work environments. Yet if these individuals are cranks, they are in respectable company.
[30] Qobs?] Yet it seems unrealistic, bordering on the surreal, to imagine
underclass blacks with their gold chains, limping walk, obscene language, and
arsenal of weapons doing nine-to-five jobs at Procter and Gamble or the State
Department. Many of these young men seem lacking in the most basic skills
required for steady employment: punctuality, dependability, willingness to perform routine tasks, acceptance of authority. Moreover studies show that even
when jobs are available, many young blacks refuse them, apparently on the
grounds that the jobs don't pay enough or that crime is more profitable.
(pp. 504-5)
[31]With some discomfort, we see that there is some truth to the historical
stereotype of the black male stud, or, at least in the case of the black underclass,
what used to be a stereotype now contains an ingredient of truth. (p. 517)
These passages give a representative impression of the various strategies
of negative other-presentation employed by D'Souza. Person descriptions of
black activista and Afrocentrists draw on familiar racist stereotypes about
conspicuous dress and lavish lifestyles. Blacks are poor and so also their
leaders should dress soberly, and at least not more conspicuously than'Nyé
do. Cultural difference is here interpreted as cultural deviance, if not as a
lack of solidarity with the black underclass. And whereas the middle class is
described as living conspicuously, underclass youths similarly are characterized in terms of the street counterpart of deviant conspicuousness (gold
chains, limping walk, obscene language, etc.) (example 30), described in
such a way as to legitimize that they are not being hired.
Thus, black young men from the 'underclass' are seen to violate all basic
values of the conservative ideology: 'punctuality, dependability, willingness
to perform routine tasks, acceptance of authority' (example 29). No wonder
they get no jobs, and they are thernselves to blame for it. Black individuality
is denied when they are described as a mindless group following their
The ideology and discourse of modern racism
303
leaders (as in example 22), behaviour that is inconsistent with the dominant
white norm of individualism. If not seen as different and deviant, blacks and
their rage are characterized as a threat (example 23). D'Souza's conclusion
is to admit that these may be stereotypes (e.g. about the 'black male studD,
but then accepts the grain of truth hypothesis to prove he must be right about
his generalizations (example 31). His analysis thus is reduced to creating the
familiar racist stereotype of the'tad nigger' (p. 524), who is portrayed as the
'menace of society'. For young black women, as we shall see in more detall
below, the stereotype is similarly predictable: they have too many babies, at
a too early age, are unmarried, and thus contribute to the lastardization' of
America.
Social problems of the black ghetto are hyperbolically characterized in
tercos of a 'civilizational breakdown' (example 25) or as 'catastrophic
cultural changé (example 26). Having to be on welfare is negatively
represented and blamed on the victims by expressions such as 'parasitic
reliance on government provision' (example 25). In other words, blacks are
lazy parasites who live out of 'our' pockets. Being ill-prepared for university
study because of bad schooling, is similarly blamed on blacks themselves,
also in tercos of aggression. Based on a few examples of some blacks who
see such achievement as 'acting white', D'Souza concludes that (ail?) blacks
share a lostility to academic achievement'. Black behaviour is interpreted
with formal style expressions such as 'dysfunctional' or less formally as
'destructive' (example 27), whereas black culture is said to have 'a vicious,
self-defeating, and repellent undersidé . Black people who have lost patience
over everyday racism, and developed a standing rage against whitedominated institutions, are deemed to be 'mentally unstable' (example 29),
so that whites seem to have a good reason not to want to hire 'such
people'.
Black women
Black women constitute a special target for D'Souza's diagnosis of black
'pathology'. Their double jeopardy when it comes to discrimination and
prejudices is clearly illustrated by D'Souza's derogatory discourse itsel£
[32] Perhaps the most serious of African-American pathologies — no less serious
than violence — is the routinization of illegitimacy as a way of life. The
bastardization of black America is confirmed by the fact that nearly 70 percent of
young black children borra in the United States today are illegitimate, compared to
22 percent of white children. More than 50 percent of black households are
headed by women. Almost 95 percent of black teen mothers are unmarried,
compared to 55 percent of their white peers. (p. 515)
Note the usual hyperboles, here further emphasized by the phrase 'perhaps
the most serious'. For outsiders of the conservative ideology, it may seem
preposterous to assume that the phenomenon of mothers who decide not to
marry has become a threat bigger than violence, and at the top of the list of
the 'pathologies' D'Souza attributes to the African-American community.
They might conclude that if that is the main problem facing the USA and
304
Discourse
Western civilization, apparently D'Souza is incapable of sensible judgements about what the real social problems are that affect US and Western
societies.
They might look at the figures and recall that many, also wealthy and
prosperous societies (like the Scandinavian countries) have similar percentages of unmarried mothers, and that most of these mothers are doing very
well, thank you. They might wonder about the functions in a book of the
1990s of old-fashioned terms such as 'illegitimaté and especially 'bastardizatior , other than to bluntly derogate, criminalize and marginalize black
women and children.
Staying within the familiar 'number gamé rhetoric, suggesting scientific
credibility, more realistic observers might again look at the figures (assuming they are correct and not themselves very selectively framed, as is usual
with statistics), and wonder about the fairly high percentages among white
women. Do these also contribute to the 'bastardizatioñ of the USA? And
how come (as D'Souza does not say) that all these percentages, in most
Western countries, are rising? Might there be a cultural change in family
values that attaches less importance to being married, and are AfricanAmerican (like Caribbean) families simply more advanced in this cultural
change, for example, by attributing more value to the leading role of
women?
Or they might ask about one of the other causes of one-parent families,
not in the stereotypical terms of irresponsible black fathers who act like
'studs' (as D'Souza so delicately uses old racist stereotypes), but in tercos of
the broad social marginali7ation of poor black men in white America.
D'Souza is even cynical when he openly legitimizes 'rational' discrimination of such young men.
Indeed, as is also clear in example 32, within the conservative framework
of D'Souzá s attitudes, racism and sexism are closely related. That such
rhetoric is not altogether without effect may be concluded from recent
decisions by local, state and national governments in the USA to dramatically reduce social welfare for poor families, a policy of which young black
women will take the brunt. In this respect, D'Souzá s book and its discourse
is not merely an innocent conservative and racist-sexist diatribe against
black women and men, or against liberals who prefer to deal with social
problems as such and not as incriminating pathologies.
In the anea of race relations, ideologies and their discourses, and even the
details of their rhetoric, may be very dangerous. They may further marginalize millions of poor black women, children and men, driving them
further finto the 'pathologies' selectively and hyperbolically attributed to
them especially in order to highlight white cultural supremacy.
It is not surprising that most African-Americans and many white liberals
agree that this kind of 'respectablé racism of cultural conservatives is more
insidious than the blatant irrational kind of the old days. They will agree
with D'Souza on one point, namely, that racism is not an 'irrational
antipathy'. His elite racism is, indeed, a deliberate, explicit and very rational
The ideology and discourse of modern racism
305
attempt to inferiorize blacks and thus to exacerbate racial inequality in the
USA.
In sum, middle- and lower-class blacks, women and men, are negatively
portrayed lexically and rhetorically in terms of cultural deviance and threat,
as violating norms and values of white America, in such a way that the
'civilizational breakdowñ of their community is fully blamed on them.
Where sober situation descriptions of social problems do not suffice,
apocalyptic hyperboles about the'breakdown of civilization' or'catastrophic cultural changé are used. Metaphors are borrowed, as usually in
such a case, from the domain of threatening animals or plants: blacks on
welfare are parasites. D'Souza is aware of his negative style when admitting
'frankness' and when boldly stating that we can no longer'euphemizé
(example 28) our language a familiar disclaimer when whites engage in
derogatives against blacks. We have earlier seen that although AfricanAmericans are the main target of D'Souza's book, also other non-Europeans
(except Asians) may share in accusations of barbarism, primitivism, threat,
deviance or lacking civilization. Of course, in the contemporary world,
Muslims are a preferred target:
[33]Muslims in the United States should be allowed to practice their religion but
not to the point where it threatens the religious freedom of others, as through the
practice ofjihad against non-Muslims. (p. 548)
Thus, historically, the others were described as 'savages', and one would
expect D'Souza to take some distance, but his own style is simply a
contemporary continuation of the old style of racist ethnocentrism.
Of course if there are bad others, there must be good others that serve as
the Good Example, and at the same time serve as evidence that 'we' are not
racist. That role, in the USA, is now played by the Asians, who have become
the model-minorities whose commercial and academic success is often used
to shame African-Arnericans or Latinos. So much so, that since they often
out-perform whites in universities, measures have been considered to limit
their numbers, so as to give the poor whites a chance. Of course, D'Souza
does not discuss such developments of reverse affirmative action. Por him,
Asians serve especially to marginalize blacks and to discredit the argument
that discrimination is still a major factor in the situation of AfricanAmericans:
[34]By proving that upward mobility and social acceptance do not depend on the
absence of racially distinguishing features, Asians have unwittingly yet powerfully challenged the attribution of minority failure to discrimination by the
majority. Many liberals are having trouble providing a full answer to the awkward
question: Why can't an African-American be more like an Asian?'
This passage has several interesting presuppositions, such as that Asians
actually are upwardly mobile, and that they are socially discriminated
against like the blacks. Implicit is the argument that Asians and African-
306
Discourse
Americans live in the same socio-economic circumstances, and have the
same start position.
Interestingly, if whites are doing worse than Asians, the question might of
course be raised about white pathologies that cause such a lag, and why also
an Euro-American cannot be more like an Asian. Many other comparisons
come to mind, such as why Americans cannot be more like Europeans when
it comes to rights of workers and social provisions, and why despite such
European 'pathologies', and despite Euro-racism, there are no ghettos in
Europe comparable to those in the USA. In other words, liberals may have
many more awkward questions than D'Souza will ever be able to answer.
Probably the only sensible answer is the one he himself provides at the end
of the book:
[35]No cace has a monopoly on achievement. (p. 472)
If 'mixed' groups have significantly contributed to the economy and the
culture of the world, as D'Souza claims, then one might wonder why he
denies such a contribution to the typically 'mixed' African-Americans. In
the full 724 pages of the book he does not once even try to assess such
contributions. That is, he can only see African-Americans in light of his
racist ideology.
Ideological enemies
Although his social enemies bear the brunt of his discursive attack, also his
ideological enemies, the Soasian relativists' are not exactly described in
positive terms, as may of course be expected from an enemy. First of all,
their theories are derogated in terms of a 'deep rooted ideology' (p. 527), a
description D'Souza hardly uses of his own ideas and those of fellow
conservatives. And because the Boasians criticized'American customs and
mores' but refrained from criticizing other cultures, they are accused of
using a'clouble standard' (p. 155). Relativists have allegedly caused the
'contemporary crisis' and do not 'allow social progress', because they treat
all cultures as equal. Relativists do not carry out research, and do not attach
conclusions to their findings, but they 'dictate' their opinions:
[36]Cultural relativism dictates that non-Western cultures be considered victims
of Western oppression: of colonialism, imperialism, racism, and so on. (p. 358)
In the same way, the relativists are seen as the source of the legal
'doctrine' of proportional representation, which implies that such representation for D'Souza is not a democratic'light'. In a more hyperbolic way, the
ideas of the others are not merely derogated as an 'ideology', but its
adherents have a 'fanatical commitment' to such an ideology (p. 530). This
lexical association with religious fundamentalism is further metaphorically
emphasized by representing liberal ideas as the innocent lamb being butchered by the relativists: 'Fundamental liberal principles are being sacrificed at
the altar of cultural relativism' (p. 530). And as we have seen, also the
metaphor of threatening forms of life will typically be employed here:
The ideology and discourse of modern racism
307
[37]Relativism has become a kind of virus, attacking the immune systems of
institutional legitimacy and public decency. (p. 532)
Apart from calling relativists 'dogmatic' as is normally the case for
ideological opponents, another useful form of negative other-presentation is
to represent the ideological enemy metaphorically as tricksters:
[38]Multicultural activists rely on the sleight-of-hand in which 'I cannot know
becomes 'I cannot judgé which becomes 'I know that we are all equal'. A
skeptical confession of ignorance mysteriously becomes a dogmatic assertion of
cultural egalitarianism. (p. 383)
In other words, such scholars are not scholars at all, but 'activists', who
hide their ignorance behind ideological dogmatism in order to avoid judging
others. Thus, relativista are routinely accused of being 'blind' to the facts,
for example about alleged 'black racism' (p. 88). Obviously, since D'Souza
does claim to know, he also has the right to judge about the black
community, as his book amply shows. Such a rejection of liberal scholars
may in fact extend to scholarship and academia in general, a well-known
feature of US conservatives, as also has become obvious in the debate about
multiculturalism and political correctness:
[39]... no good is achieved by dressing these pathologies in sociological cant,
complete with the familiar vocabulary of disadvantage and holding society to
account. (p. 486)
The derogatory label 'sociological cant', familiar also from tabloid
reactions against anti-racist academics in the UK, 6 apart from expressing the
conservative and anti-relativist attitudes of D'Souza, also may be interpreted
as a move in a broader 'commonsensé strategy, in which ethnic relations
and racism should rather be examined in populist tercos (of course those of
the conservatives). S ociologists might persuasively argue and prove that the
conservative solutions offered by D'Souza will only exacerbate the social
misery of many inner-city blacks, as neo-liberal policies elsewhere tend to
exacerbate poverty, and make the rich richer. Or they might show (as they
have done, but all those studies about modern racism are either ignored or
rejected) that racism is still prevalent in the USA, and that it is still a major
factor in explaining the many social and economic gaps between black and
white.
No wonder that such sociologists are simply dismissed by D'Souza. Their
insights might be dangerous for his conservative analysis of the situation, as
is the case for a book by Joe Feagin and Melvin Sikes, Living with Racism:
the authors, who are not called 'distinguished scholars' or simply 'sociologists', but 'activist scholars' (p. 491). Thus, any serious evidence of
everyday racism is simply marginalized or ignored and negatively presented
by derogating their authors as 'activists', and hence as not being 'objectivé .
The same is true for those black professionals whose experiences with
everyday racism are presented in. Feagin and Sikes's book. The powerless
rage and discourse of these black women and men are clinically diagnosed
308
Discourse
by D'Souza as a form of 'mental instability' (p. 492), another well-known
strategy to problematize and marginalize the others. Thus, D'Souza brings to
bear all discursive devices to ward off any inconsistencies with his claim
that racism in the USA is no longer a problem. Thus, prominent black
scholars who are not members of his league of conservative blacks, may
simply be ridiculed, as we have seen before, as is the case for Cornell West,
whose
[40]solutions are a quixotic combination of watered-down Marxism, radical
feminism, and homosexual rights advocacy, none of which offers any realistic
hope for ameliorating black pathologies. (p. 520)
One needs few explicit discourse theories to analyse the derogatory labels
of such a passage, and inferring who the various bad others are in D'Souza's
uníverse. Ironically, when blacks do make it in white America, and become
prominent professors, they still do not seem to escape the 'pathologies' diagnosed by D'Souza, unless of course they espouse, as some do,
D'Souza's ideology. His condemnation of the African-American community, thus, is one of principie, and not one of fact or generally shared
criteria.
This, then, is the hallmark of the conservative and racist ideologies
promoted by D'Souza, by whose standards blacks (with some windowdressing exceptions) are inferior, whether they are poor mothers in the
ghetto, or prominent scholars at Princeton, especially when they write books,
as West did, aptly called Race Matters. Indeed, many passages of D'Souza's
book about the 'inferiority' of blacks show how valid are conclusions like
those of West, as is the case in the following combination of the metaphors
of 'incantatioñ and 'demons' intended to ridicule scholars and to deny
racism:
[41]The charge of racism becomes a kind of incantation intended to ward off the
demons of black inferiority. (p. 529)
The most fatal accusation levelled against scholars is that their relativism
ends up denying the possibility of truth' (p. 384), a truth which of course
D'Souza claims to uncover in his book. Since relativists have been shown to
be about black racism, D'Souza sees it as his task to enlighten his
readers about what he does see. The others, relativists and black intellectuals, are thus accused of 'moral paralysis' (p. 520). This example also
shows the close relation between group ideologies and the self-attribution of
truth in social representations, as well as the relation between ideology, truth
and the moral order.
Thus, D'Souza sucos up the derogation of his ideological enemies by
accusing them of condoning if not promoting larbarism', and hence of
being enemies of the nation, if not beyond the pule of Western civilization in
general:
[42]By refusing to acknowledge that one culture is better than another — by
erasing the distinction between barbarism and civilization — cultural relativism
cruelly inhibits the nation from ideritifyingand working to ameliorate pathologies
that are destroying the life chances of millions of African-Americans. (p. 528)
The ideology and discourse of modern racism
309
Of course, the apparent empathy expressed by the phrase'the life chances
of millions of African-Americans' is merely a device of impression management of the harsh racist ideology underlying such passages, namely, that of
white cultural supremacy.
Positive self-presentation
Contrary to this apocalyptic picture of blacks and their culture, as well as all
other others, the description of Us, whites, Western civilization, European
explorers, and even racist scientists, is squarely positive or mitigated, when
they are obviously engaged in loathsome action. Early racism itself is
described as 'a scientific ideology to explain large differences in civilizational development that could not be explained by environment' (p. 22)
and thus given at least some scientific legitimacy, as also the descriptors
leading scientists' and 'progressive thinkers' (p. 120) suggest. Ethnocentrism is merely an 'intense preference for oné s own group' (p. 35). Of
Western technology spreading over the world, only the positive 'comforts'
and a 'cosmopolitan awareness' are mentioned, and not a single of the
negative consequences.
That such a positive evaluation is not limited to historical racism but still
applies today may also be seen in the many ways present forms of racism or
ethnocentrism are euphemized, mitigated, excused or explained away, if not
plainly denied:
[43]... it is entirely possible that prejudices might be prudent, stereotypes may
contain elements of truth, and racial discrimination may be warranted under some
circumstances. (p. 120)
That is, except when used in connection with Ilack racism', such 'white
racism' (a term seldom used by D'Souza, of course), is consistently being
put between quotes, or in accusatory contexts, except in its extremist,
irrational forms (largely deemed to be occurring only in the past). Racism,
thus, becomes something the others Invent' (p. 238) or that is Imaginary'.
If recognized at all, it is mitigated, relegated to the past or rationalized as
'natural' ethnocentric 'preference for one' s own group'. Or in more scholarly terms it can be denied by claiming that 'It is impossible to answer the
question of how much racism exists in the United States because nobody
knows how to measure racism and no unit exists for calibrating such
measurements' (p. 276). Such scientific style merely functions to impress or
persuade those who have no knowledge of scholarly studies of racism.
The denial or mitigation of racism, thus, not only serves within the
strategy of positive self-presentation, but at the same time may be used, by
turning the accusation around, to blame the blacks, as in the following
passage, which deserves to be quoted in full, and analysed in some more
detall, because it expresses many of the beliefs of D'Souza about racism:
[44]Sometimes racism is all too real, but it is bad enough to endure real racism
without having to suffer imaginary racism as well. Racism has become the opiate
of many middle-class blacks. For society, promiscuous charges of racism are
310
Discourse
dangerous because they undermine the credibility of the charge and make it more
difficult to identífy real racists. For blacks, the risk of exaggerated and falce
charges of racism is that they divert attention from the possibilities of the present
and the future. Excessive charges of racism set up a battle with an adversary who
sometimes does not exist.... Once again, racism becomes the culprit now
accused of having taken an even subtler and more insidious shape. (p. 4875
This passage begins with a familiar strategic move of positive selfpresentation, namely, a so-called apparent concession. This concession is
apparent because in the rest of the passage, and the rest of the book, hardly
any white racism is being detailed. Secondly, 'it is bad enough .. i s
another move, this time of apparent empathy, which I call 'apparent'
because D'Souzá s book is not at all empathetic with the victims of racism.
Both moves here serve as introductions to the reversal, which is introduced
with the claim that blacks not only imagine racism but that it even serves the
interests of the black middle class. Thus, the victims of racism are not just
blamed for it, they are even accused of enjoying it, as the use of the
metaphor of addiction ('opiate') shows, a charge that is of course consistent
with the dominant prejudice about blacks as selling or being 'on drugs'.
In the next sentence, another dimension of black 'pathologies' is
expressed, namely, the well-known element of 'threat' to society. The same
sentence further emphasizes ('promiscuous') the well-known counteraccusation that racism is only in the mind of the accusers. Note that
'promiscuous' here ties in with the other 'pathology' of the black community, and of black women in particular, namely, sexual promiscuity. Having
conceded'somé racism, it needs to be identified, and so it is attributed to
the 'real racists', who were earlier defined as those who believe in biological
racial superiority and 'irrationally' discriminate blacks. Such usage implies,
of course, that most of white society is not racist, as the underlyíng attitude
has shown.
Similarly, where racism is denied, mitigated or safely attributed to 'real
racists' (a move that may be called blame transfer, which is typical for elite
racism), the opposite is true for the accusations of racism, which are called
'excessivé, thus enhancing the contrast between Us and Them. The'imaginary' nature of racism is further emphasized by accusing blacks of paranoia,
of imagining non-existent adversaries, thus bringing failing mental health
into the picture, as we have seen earlier, for those whose 'rage' cannot be
understood. Finally, this passage ridícules the accusation that modem racism
is more subtle and insidious than the old one, and thus also rejects that
charge. In sum, D'Souza uses several discursive devices to persuasively
formulate his attitude about racism, and these devices all background white
racism and foreground black pathologies (imagining things, the use of an
opiate, excessiveness, paranoia).
Whereas throughout the book, thus, the rosy picture of Western Civilization, including the abolition of slavery ('Abolition constitutes one of the
greatest moral achievements of Western civilizatioñ, p. 112), is highlighted,
not a single word is used to describe the negatíve dimension of Our culture.
The ideology and discourse of modern racism
311
Slavery? No, that was not our invention. Moreover, D'Souza claims, 'the
American slave was treated like property, which is to say, pretty well'
(p. 91). And in any case, slavery cañ t be blamed anymore for black
'pathologies' today. Colonialism? No, because colonialism only brought
progress and put an end to barbarism and primitiveness, it was a'bold
intellectual enterprise to dispel ignorance' (p. 121).
Only sometimes do we find a very tentative disclaimer about Our failure,
for example as in the following (duly euphemized) apparent concession
about 'mixed motives':
[45]Whatever their shortcomings and mixed motives, the Europeans who
voyaged abroad were the historical instruments of a major world transformation.
(p. 49)
In other passages, we find plain denials, sometimes accompanied by a
complete reversal of the charge, as in example 38:
[46]These Europeans did not approach Asia, Africa and the Americas with
hostile intentions. (p. 48)
[47]What distinguished Western colonialism was neither occupation nor brutality
but a countervailing philosophy of rights that is unique in human history.
p. 354)
Thus, our Western civilization is described in tercos of 'powerful ideas'
and 'progress' (p. 50), as 'moving ahead' while 'other groups' are portrayed
as 'stagnane , thus rhetorically enhancing the contrast between Us and Them.
Of course, describing Our culture in such tercos generates a bit of uneasiness, but there is a strategy to deal with that, namely, to blame the others for
not doing the same:
[48]Since contemporary scholars do not hke to think of cultures as superior or
inferior, advanced or backward, the very subjects of primitivism and progress,
development and underdevelopment, frequently generate discomfort and even
indignation. (p. 55)
No words, of course, about other highlights of Our culture, such as the
Holocaust, or pollution, or world wars, or the atoro bomb, to narre only a
few. Thus, the underlying ideology of Western supremacy also shows in the
one-dirnensional attitudes and finally in the lexical and rhetorical forros of
selectively positive or euphemistic self-descriptions. The contrast with the
black 'barbarians' could not be greater, as the polarized structure of intergroup ideologies predicts.
Conclusions
The aim of our partial analysis of some passages of D'Souza's book The End
of Racism is to see ideologies at work. I examined some of their propositional contents, their structures, and how they control specific attitudes about
a number of issues. Finally, I showed how such underlying social representations also control many properties of discourse.
312
Discourse
The analysis has shown how a specific ideological text of an individual
author combines influences from several ideologies, within a broader framework of cultural conservatism. Thus, we find a combination of ethnocentric,
racist, sexist, anti-relativist and neo-liberal ideologies in the construction of
complex attitudes about African-Americans, racism and anti-racism, multiculturalism and Afrocentrism and other attitudes. The conservative framework and its propositions and underlying values assign coherence to these
attitudes and show how they are mutually related. Ideological polarization
has been shown for the representation of blacks and whites, barbarians and
the civilized, realists and relativists, Us and Them. Ideological schemata
organize such propositions in terms of what They are, what they typically (if
not stereotypically) do, what their aims and values are, how Us and Them
are related (namely, as superior and inferior) or what their resources are.
A succinct and informal discourse analysis has further detailed this overall
ideological analysis, and highlighted the social and political functions of this
text, and how its discursive devices are tuned to the persuasive communication of the ideology of modern racism. Group polarization is thus expressed
and enhanced by a series of well-known devices that emphasize how bad
They are and how good We are, or that mitigate their success and our
failures. Overall, derogatory lexical style, rhetorical devices (such as metaphor and hyperbole), local semantic moves of denial and apparent concession, the rhetoric of factuality by the use of (selective) statistics, and many
other features of this text can be described and explained on the basis of the
underlying ideologies and prejudiced attitudes.
It was concluded that such racist ideologies and the discourses that convey
or reinforce them are not merely academic exercises nor food for media
debates. They explicitly formulare and propose harsh social policies. They
are read by influential conservative politicians and other elites, eagerly
accepted as a scientific legitimation of racial bigotry, prejudice and the
marginalization of blacks, and actually used as the basis for racist policies
that contribute to ethnic and racial inequality in the USA.
29
Conclusions
Instead of a lengthy discussion of the findings of this theoretical study, I
shall merely list its major conclusions in the forro of brief statements.
General
1 Within the vast field of the study of ideology, a multidisciplinary
theory is needed to account for the nature, the structures and the functions of
ideology.
2 In this study, this multidisciplinary approach is represented by an
analysis of ideology in terms of the 'triangle' of (social) cognition, society
and discourse. This complex disciplinary basis is necessary to avoid
reduction. Especially lacking in earlier work is insight into the sociocognitive nature and functions of ideologies, and how these are related to
their expression and reproduction in discourse.
3 Many of the traditional approaches to ideology are rather of a
philosophical than of a systeniatic, analytical and theoretical nature. The
confused and often vague nature of traditional ideology studies is also due to
the repetition and uncritical acceptance of a number of standard concepts of
studies of ideology in the past. A typical example is the notion of 'false
consciousness'. Perhaps the most promising work on ideology is currently
especially done in the study of political cognition and social representations.
4 In a general and abstract sense, ideologies are conceived of as the
interface between fundamental properties (e.g. interests, goals) of social
groups and the shared, social cognitions of their members.
5 Compared with commonsense and traditional Marxist or other sociopolitical definitions, ideologies are here defined in a general, non-pejorative
sense (and not necessarily as false, or distorted ideas).
Socio-cognitive analysis of ideologies
6 The cognitive analysis of ideologies does not imply that ideologies are
individual or only mental. They are both mental and social, and also their
mental properties are socially acquired, shared and changed.
7 Ideologies are most generally defined as systems of beliefs, especially
in political psychology. However, it was argued that there are many types of
314
Conclusions
belief, many of which are not Ideological'. So, a theory of ideology needs to
focus on specific, ideological beliefs.
8 The traditional distinction between episodic and semantic memory is
used to distinguish between personal beliefs, on the one hand, and social
beliefs or social representations, on the other hand. Ideologies are of the
latter kind, and hence first (and as yet incompletely) defined as shared,
social beliefs of (specific) social groups.
9 Since on the other hand there are also several types of socially shared
beliefs (knowledge, actitudes, norms, values, etc.) some of which are not
ideological, it is further proposed that ideologies are the general, abstract
beliefs that underlie (other) social representations. In that respect, they are
like the basic axioms of the system of social representations shared by a
group.
10 Ideologies are not arbitrary lists of propositions, but organized by
specific social categories that constitute an ideology-schema, such as Membership, Activities, Goals, Values, Position or Resources. These categories
are the cognitive (re)construction of the basic social criteria for groups.
Cognitively, this schema functions also as the self-schema of the group,
defining its social identity and interests.
11 Ideologies also may have other structural characteristics, such as
those of group polarization (Us versus Them).
12 Ideologies are the basic social beliefs of specific groups, but themselves rooted in the general beliefs (knowledge, opinions, values, truth
criteria, etc.) of whole societies or cultures. This allows the very understanding, communication and interaction between (members of) different
groups.
13 Ideologies, as social representations, are generally assumed to be at
least coherent. Such coherence explains the frequently observed coherence
and continuity of ideological opinions, practices and discourses among
different social members and in different situations.
14 Ideological coherence does not imply that ideologies are always used
coherently by group members. That is, the equally frequently observed
variability of discourse or social practices monitored by ideologies, is not
due to lacking ideologies or incoherent ideologies, but by several other
factors, such as the interaction of several ideologies (and group memberships) for social members, personal experiences, and the constraints of the
situation.
15 Ideologies are coherent and complex only at the group level for
which they are defined. Depending on their social position and socialization,
different (sub)groups of social members (e.g. the ideologues) may have
different ideological expertise.
16 The main cognitivefunction of ideologies is to organize the social
representations of a group. Indirectly, that is, through more specific, domainrelevant, attitudes and knowledge they thus monitor social and personal
beliefs and ultimately the social practices and discourse based on the
latter.
Conclusions
315
17 Attitudes are here defined as socially shared complexes of the shared
opinions of social groups, and are carefully distinguished from personal
opinions.
18 A distinction is also made between factual beliefs (trae or false
knowledge) and evaluative beliefs (opinions, attitudes, ideologies) which are
based on the application of socio-cultural values.
19 The well-known problem of the relation between social knowledge
and ideology was resolved by making a distinction between (historically
variable) cultural knowledge that serves as a 'common ground' for all
(competent) members, on the one hand, and the specific knowledge of a
group (which may be called 'opinions' by members of other groups). It is
the latter kind of group knowledge that may be ideologically controlled.
Group knowledge may sometimes become general cultural knowledge and
vice versa.
20 Although ideologies may thus also control group knowledge, they
especially monitor the shared evaluative beliefs (opinions) of a group. They
are the basis of the social judgementS of groups and their members.
21 Ideologies are not defined as wrong, misguided, false, or distorted
beliefs of a group. Epistemically, whatever their truth status for the group
itself, they may be trae or false. It is not their truth value, but their cognitive
and social role (e.g. effectiveness, usefulness) in the management of thinking
and interaction that is the criterion for their evaluation.
22 The socio-cognitive notions introduced aboye explain more analytically such notions as ideas, beliefs, (false) consciousness, common sense, in
traditional studies of ideology.
23 In order to explain how socially shared representations in general,
and ideologies in particular, can be related to personal cognitions (and then
to discourse), the notion of mental model is used, for example to account for
the subjectivity of personal experiences, interpretations and representations
of discourse and action and the representation of contexts.
24 Models are the interface between the social and the personal,
between the general and the specific, between the macro and the micro. They
apply or instantiate socially shared information (knowledge, attitudes, ideologies) in relation to self, to current situations, tasks, problems, actions and
discourse. Conversely, they are the experiential basis for the generalization
of personal beliefs to social knowledge, attitudes and ideologies.
25 Models embody personal and applied social beliefs, and thus, indirectly, ideologies. It is through ideologically controlled models that ideological social practices and discourse can be produced by social members.
Social analysis of ideologies
26 Ideologies are by definition social; they are socially shared by
groups. They are only individual in their personal, contextual uses, applications or implementations by individual social members. In that respect they
are like language systems (or grammars, or discourse tales).
316
Conclusions
27 Ideologies are not generally social or cultural, but defined for specific
social groups. Not all collectivities of people form such groups, but only
those collectivities that satisfy a number of group-criteria, such as (more or
less continuous, permanent and organized) membership conditions, joint
activities, interaction, goals, norms and values, a specific position in society
and social resources, and especially shared social representations. These
precisely map onto the cognitive structures of shared ideologies. Thus,
groups constitute ideologies (and hence social identity) just as much as
ideologies constitute groups.
28 The socialfunctions of ideologies are tied to these properties of
groups. They represent group identity and interests, define group cohesion
and solidarity, and organize joint actions and interactions that optimally
realize group goals. That is, ideologies resolve the fundamental problem of
social and interactional co-ordination, namely, that, despite personal and
contextual variation, individual social actors are generally able to act as
group members, and often in the interest of the group as a whole.
29 Ideologies are especially relevant for the management of social
group relations, such as those of domination and conflict, but also those of
competition and co-operation. It is in this respect that ideologies may
function as legitimation of power abuse and inequality, on the one hand, and
as a basis for resistance, challenge, dissidence and change on the other
hand.
30 Given the definition of ideology in terms of social groups, they are
not limited to dominant groups. Such would unduly restrict the notion and
make it theoretically much less interesting. For one thing, it would prevent
an ideological analysis of dominated groups and practices of resistance.
31 Because of their preferential access to, and control over, public
discourse, and especially of the media and education, various elites have a
special role in the formulation and reproduction of ideologies. Although
ideological reproduction is both top-down and bottom-up, this suggests that
a relatively small number of 'symbolic' elites (writers, thinkers, politicians,
scholars, journalists, etc.) may exercise the special role of ideological
leaders, who preformulate and stimulate ideological debate.
32 The effective reproduction and implementation of group ideologies
often requires organization and institutionalization, typically so by ideological institutions such as those of politics, the media and education.
Ideologies and discourse
33 As described aboye, social group ideologies indirectly (and hence
non-deterministically) monitor social practices in general, and discourse in
particular, via social beliefs (knowledge, attitudes) and personal beliefs
(models).
34 Discourse has a special function in the expression, implementation
and especially the reproduction of ideologies, since it is only through
Conclusions
317
language use, discourse or communication (or other semiotic practices) that
they can be explicitly formulated. This is essential in contexts of acquisition,
argumentation, ideological conflict, persuasion and other processes in the
formation and change of ideologies.
35 Despite the fundamental role of discourse in the expression and
reproduction of ideologies, ideologies cannot be reduced to discourse. That
is, they should not be defined as statements, and their nature and structure
should not be identified with the structures of text or talk. An analytical
distinction should be made between ideologies as general, abstract, sociocognitive (mental) representations shared by a group, on the one hand, and
the specific, personal, interactional, contextualized uses of the ideology in
specific social situations by individual social members, on the other hand.
Indeed, if ideologies were to be reduced to (or identified with) discourse, it
would be impossible to explain how they can influence other social
practices.
36 An analysis of the discursive expression and reproduction of ideologies requires a detailed, systematic account of the various levels, structures,
units and strategies of text and talk, defined as communicative events. Such
an analysis should not, as was traditionally often the case, be limited to a
vague study of the 'production of meaning'. Besides complex semantic
analysis of various types of meaning, also explicit other theories are needed
to account for these discourse structures and how they may express
underlying ideological contents and structures, for example, phonological,
graphical, syntactic, lexical, stylistic, rhetorical, schematic (e.g. argumentative, narrative), pragmatic and conversational structures.
37 Besides an account of the levels and structures of text and talk,
discourse analysis also provides a detailed analysis of the many properties of
the context, defined as the discourse-relevant structures of the social situation. Context influences discourse 'uses' (production and comprehension)
through subjective mental models of language users, that is, through context
models.
38 Ideological discourse production is a complex social and cognitive
process in which underlying mental models are mapped on discourse
structures — for example, mental models of events map on to semantic
structures, and mental models of context on to the large number of variable
discourse structures (forms, expressions, schemata, etc.). Context models
exercise the overall control of such discourse production and ensure that
discourses are socially (or indeed, ideologically) appropriate in the social
situation.
39 Ideologically based mental models as well as more general social
representations may thus be expressed or signalled at all levels of discourse
structure, that is, forms, meanings and actions. The overall strategy hereby is
in fine with ideological polarization and other structures, such as self-serving
positive self-presentation and negative other-presentation.
40 This overall strategy may be implemented by a large variety offorms
and meanings that emphasize (or mitigate) positive (or negative) properties
318
Conclusions
of the ingroup and the outgroup, respectively, for example through intonation, stress, volume, clause structure (transactivity: e.g. actives and passives), lexical selection, implicitness, presuppositions, local coherente,
overall topics, rhetorical devices (e.g. metaphors), schematic organization
(argumentation, fallacies), the selection of speech acts, and conversational
and interactional management (e.g. of politeness).
41 Conversely, in discourse comprehension and persuasion, these various discourse structures may in turra be used to influence the formation, the
contents and structures of mental models and, often indirectly, of social
representations and hence of ideologies. These strategies are generally tuned
to the formation or change of preferred models or their structures, again
under the general constraints of positive self-presentation and negative
other-presentation strategies.
42 However, ideological influence and reproduction are not merely a
function of discourse structures but also of the social context (or rather of
context models), and of the (other) mental representations of the recipients,
such as existing ideologies, attitudes, knowledge, models of experience,
current goals and personal interests, and so on. This means that ideological
influence may not always have the intended effects. Despite their group
membership, and the powerful influence of social representations, social
actors are in principie autonomous individuals, and hence largely in control
of their opinion formation and change, for example as a function of personal
interests, goals and wishes. Ideological influence, and hence reproduction,
will hence be most successful if ideologies are consistent with personal
experiences (models), if social actors have no (better) alternatives than the
proposed ideologically based models for their opinions and actions, or if
they can be manipulated to believe and prefer (misguided) information
('facts', opinions) even if it is not in their best interests.
43 A sample analysis of a book about race relations in the USA shows
(a) that social attitudes, personal opinions, event models and discourse may
exhibir an interaction between various ideologies, (b) that conservatism is
rather a'meta-ideology' than an ideology, (c) how social groups (Us and
Them) are represented in attitudes and discourse, (d) how at many levels of
text and by many devices ingroups are presented positively and outgroups
presented negatively, and (e) how ideological discourse is (made) sociopolitically relevant in times of (real of imaginary) social crisis, as a means to
confirm group dominance and to legitimate inequality.
Limitations and prospecta
The outline of the theory of ideology presented in this book and summarized
aboye is just that: an outline. Yet, it tries to offer a comprehensive
framework for detailed theoretical and empirical studies of ideology.
As suggested, such studies need to be multidisciplinary. One of the main
limitations of traditional studies was that they ignored systematic and
Conclusions
319
analytical insights from other theories and disciplines. Indeed, I have shown
that the cognitive and especially the discursive dimensions of the theories
were hardly developed.
This meant that the classical, socio-economic approach could only be
formulated in very general, abstract and often vague tercos. Ideologies
empirically only 'show' in social interaction and discourse, as well as in
their organizational and institutional structures, and hence they need to be
empirically studied at those levels.
Moreover, a social explanation of ideological interaction and discourse is
unable to relate social structure to interaction and discourse structure, and
needs a cognitive interface. This cognitive interface, however, cannot simply
and vaguely be identified with 'belief systems'. We need a much more
detailed analysis of mental representations and mental strategies in order to
understand how ideologies relate to social practices, and to discourse, and
how they are thus reproduced.
I have tried to elaborate a theory that establishes these various relations.
Obviously, many elements of the theory are not yet fully worked out. For
instance, given the predominant social nature of traditional studies, I have
only paid attention to some aspects of the role of social interaction and social
structure in the formation, functions and reproduction of ideologies. I have
assumed that ideologies are by definition group-based. However, we need to
spell out in more detail under what conditions groups develop ideologies,
and indeed how ideological groups are formed. We need to pay much more
attention to the organizational and institutional dimensions of ideologies and
the ways they function and are reproduced in society. Ideological conflicts
need to be analysed in detail in order to understand the role of ideologies in
such conflicts.
Similarly, despite the relative detail provided for some of the cognitive
aspects of ideologies, there are many blank spots on the mental map of the
structures, contents, organization and functions of ideologies. We have
provisionally assumed an ideological schema based on social group selfschemata, but such a schema may be too specific, and not adequate for more
general and 'universalise ideologies (such as religions and complex political
ideologies). We need to know much more about the ideological control of
the (structures) of other social representations, such as attitudes and knowledge. We only have tentative ideas about the relations between (personal,
subjective) models of experience, and the socially shared representations of
the group. Indeed, how and under what conditions are mental representations
personal, and when are they socially 'shared' or Inowñ in the first place?
We know as yet very little about the internal organization of mental models
and how they embody (ideologically based or other) knowledge and opinions. And fmally, we have only vague ideas about the precise relations
between models and social representations on the one hand, and discourse
structures or social practices on the other hand.
Finally, only a beginning has been made to an explicit analysis of those
structures of text and ta1k that systematically express, convey, signal,
320
Conclusions
communicate or influence underlying ideologies. Although in principie all or
most discourse structures may be so used, it may very well be that some do
so more typically or more effectively. Much empirical work will be needed
to show how some groups use (and abuse) discourse in very specific ways.
Indeed, what kind of ideological discourse is typical for what groups, what
are its properties, and how is it in turra socially and institutionally embedded? How are ideologies discursively expressed and reproduced in such
important social domains as politics, the media and education?
In sum, there are many more questions left open than answered in this
book. It should therefore rather be seen as a sketch for a research programme
than as a complete theory of ideology. As strongly suggested, such a
research programme can only be carried out successfully if scholars from
different disciplines (and knowledge about each other's theories and concepts) combine to elaborate the theoretical and empirical details. The
development of a fully fledged theory of ideology cannot be left only to
psychologists, or only to social scientists, or only to discourse analysts, or,
indeed, only to philosophers.
Index
abstraction, 252-4
levels, 17, 19, 24, 54, 207
access to public discourse, 173-4, 261-2
acquisition of ideologies, 4.247-8, 253
affect see emotions
affiliation, 223-4
affirmative action, 295-6
African-Americans, 277-312
derogating, 301-3
Afrocentrism, 297
alienation, 96
alliteration, 272, 273
Althusser, Louis, 2
anti-intellectualism, and 'common sense',
104-5
arguments
based on 'common sensé, 104
structure, 66, 67
attitude change, 244, 253
attitude structures, 60-4
altitudes, 25, 33, 43-8, 245, 253, 315
about affirmative action, 295-6
about Afrocentrism, 297
about multiculturalism, 296-7
about racism, 293-5
componente, 61
and context models, 213
defence of concept, 43-6
direct expression of ideological, 240
in ethnocentrism/racism, 290-9
and ideologies compared structurally, 65-7
audience sine see scope
awareness
consciousness as, 98-100
and denial, 100-1
training, 99
base-superstructure model, 2
basic beliefs, 20, 24
belief systems, 20, 313, 319
political, 3
and power abuse, 165-8
to social cognition, 126-7
use of terco, 28
beliefs, 18-21
aboutness, 21, 25
acceptability, 232
clusters see attitudes; knowledge
and cognition, 21
different kinds, 127
forms of representation, 22-6
and ideas, 15-27
ideologies as, 26-7
and knowledge, 18-19
levels of abstraction, 24
and natural language, 23, 24
problems in definition, 23-6
propositional description, 22-3
representation-relation, active/passive nature,
25
trae and false, 25, 34, 41, 42
types, 41
see also basic beliefs; general beliefs;
particular beliefs
Bewufztsein see consciousness
'biased' models, 210, 234, 241, 251
black culture, 'pathologies', 277-312
black women, D'Souza's racist discourse,
303-6
Boas, Franz, 278,281
body, versus mirad, 17-18
bottom-up legitimation, 174-6, 257
brain, neural networks, 23, 57
brain-mind problem, and consciousness, 100
categories, 57
change, possibility of ideological, 88, 94
churches,186
class discourse, 140, 179, 181
class membership, 152
classical approaches, 2, 4, 191
co-operation, 171
co-ordination problem, 165-8, 316
codes, semiotic, 197-8
coercion, 162,166
cognition, 15-134
and beliefs, 21
discourse, 235-42
society and discourse triangle, vi' 5, 11,
131, 136-7, 313
Index
cognitive, and social, 9-10, 51-2, 135-7
cognitive analysis, 27
cognitive competence, 246
cognitive dissonance, 63, 93
cognitive psychology, 26, 47, 54
cognitive science, 10, 16-17, 130
cognitivism, 10, 27
coherence
as condition of continuity and reproduction,
94, 314
and consistency, 91-2
global, 206-7
local, 206-7, 269-70
collective action, 141, 143
legitimation as, 256
collective memory, and group identity, 123
collectivities, 29-30
'common knowledge' of culture, 37-41, 50-1,
86-7
'common sense', 98, 102-7
dimensions, 103-6
ideologies as, 106-7
meanings, 102-3
communication see discourse
communicative acts, intentional, 216-18
communicative events, versus verbal products,
193-4
communism, 2, 3
competence, 194
abstract versus practica!, 54-5
see also cognitive competence; social
competence
competition, 11, 72,170
conflict, and struggle, 145, 168-70
conflicting ideologies see oppositional
ideologies
conformity, 148-9
connectionist approach, 23, 57-8
consciousness, 96-101
as awareness, 98-100
raising, 99, 261
consensus, 3, 36, 163, 167, 181, 190, 274
power of, 114-15
consent, 163
manufacture of, 274-5
conservatism
discourse of Us, 277-312
as meta-ideology, 284-7, 312, 318
consistency, 88, 90-5
and coherence, 91-2
versus variation, 90
constructivist approach, 25
contents, 72-3
context, 211-27, 317
defined, 211
367
dimensions, 214-27
in reproduction of ideologies, 230-3
sensitivity of ideologies to, 55-6
context models, 82-3, 86-7,133, 212-14, 236,
317
expression, 241
and mapping of ideologies onto
communicative events, 86-7
and opinions, 249,252,253-4
control
ideological, 209, 227, 234, 260, 262
issues, 161-4
conversation analysis, 53, 54,200,212,217,
273-4
copies, 195
counter-ideologies, 130,167,261
credibility, 244, 252, 265-6, 274
'crisis', civilizational (lYSouza), 282-3
critical approach, 11-12
cultural beliefs
and 'common sense', 106-7
versus group beliefs, 36-41
cultural knowledge, 37-41, 50-1, 86-7
and group knowledge, 60-1
'cultural racism', 278
cultural relativism, 279, 306-7
cultural studies, 3
culture
role in development of ideologies, 135
and social identity, 124
definitions, 1,113-14
delegitimation, 258-62
denial, and awareness, 100-1
description, modes, 53-5, 80
discourse, 6,26, 191-312
concept, 193-8
extended and restricted notions of, 197
from cognition to, 235-42
and ideology, 191-9, 316-18
indirect links to ideologies, 85-8,132
legitimation and delegitimation, 260-2
and presupposed meanings, 31
relevance, 191-9
role in reproduction of ideologies, 5, 6,
230-4, 316-17
roles and norms, 6
society and cognition triangle, vii, 5,11,
131,136-7, 313
see also intergroup discourse; intragroup
discourse
discourse analysis, viii, 6, 198-9, 255
discourse control, and ideology, 210
discourse formation, 196
discourse processing, cognitive theory, 235-42
368
discourse production, 236-9, 317
formulation module, 238-9
pragmatic module, 236-7
semantic module, 237-8
discourse structures, 200-10, 271
ideological see ideological discourse
structures
in reproduction of ideologies, 233-4
discourse studies see discourse analysis
discourse syntax, coded for ideological
functions, 207-8
discursive events
circumstances, 219-20
date and time, 219
location, 219
discursive 'repertoires', 43, 45
dissidence, ideological, 183-4
domain, 214-15
dominance, 258, 274-5
dominant classes, 179
dominant groups, 259
dominant ideology, 97, 98, 102, 108, 130,
179-85, 255
'dominated groups, 181-3
domination, 2, 68, 72,140
and power, 161-4
doxa, and episteme, 19, 34
D'Souza, Dinesh, viii, 277-312
Durkheim, Émile, 2,108
economic determinism, 3
education system, 186-7
effects, ideological, 244-5
elite discourse, 175, 262
elite groups, 3, 105, 108
elite ideologies, 97-8
and strategies of control, 184-5
elite pówer, and professional knowledge, 50
elite racism, 176-8
elites, 140, 149, 172-8, 259-60, 316
embeddedness, 11, 192
emotions, 20-1, 62
and social identfty, 122
End ofRacism, The (D'Souza), viii, 277-312
Engels, Friedrich, 2, 96, 108, 179
Enlightenment, the, 15
entailment, 207
episodic memory, 29-31, 33, 79, 81,133,
212-14, 239-40, 244, 250, 314
episteme, and doxa, 19, 34
epistemic order of society, 34
epistemology, 35-6,109-10
errors, 54
ethnocentrism/racism, 287-90, 309-10
ethnography of speaking, 198, 212
Index
ethnomethodology, 102
euphemisms, 270, 272, 273, 309
evaluation entena, versus truth entena, 34-6,
41
evaluations, 19
organizing, 59-60
evaluative beliefs 85-6; see also opinions
event models, 79, 81, 83, 87, 133, 249
expression, 240-1
event schemata, 81-2
event tokens, 84
event types, 84
exclusion, and inclusion, 159-60, 161
experience models, 80, 81
expertise, 262, 314
explication, 106
explicitness
and implicitness, 207
implicitness versus, 268-9
expression of ideologies, 8, 90,239-42
direct, 192-3, 239, 263-4
in discourse structures, 200-10
indirect, 192-3, 263-4
instantiated indirect, 239-40
stability, 56
expressions of belief, 22-3, 24, 26, 42
factual beliefs see knowledge
false, meanings of, 97; see also truth/falsity
debate
false consciousness, 15, 73, 87, 96-8, 108,
260-1
families, 186
Feagin, Joe, 307
figures of style see rhetorical structures
Foucault, Michel, 193
Frankfurt School, 11
Fred erickson, George M., 277
freedom, ideological claims for, 162
functi ons of ideologies, 3, 7-8, 314
Galileo Galilei, 28
general beliefs, versus particular beliefs, 31-2,
41
generalization, 229,252-4,270
generative grammar analogy, 47, 57
genres, 196, 215-16, 264-5
functions, 216.218
goal ideology, 69, 70, 218
grammar analogy, 32, 90,148,150
Gramsci, Antonio, 2, 3, 102, 140
graphics, 201, 273
group beliefs
ideologies and, 37
versus cultural beliefs, 36-41
group categories, and membership, 151-60
Index
group cohesion, 11
group conflict, 67-72
group identity, 70, 71-2
and collective memory, 123
ideology as, 120-2
and personal identity, 118, 119-25
group ideology, 91-2, 120-1
group interests, 70,115
group knowledge, and group-specific entena,
38
group membership, 143
and personal identity, 120
group polarization, 67-72,161, 314
group relations, 161-71, 316
role in development of ideologies, 135,
161-71
group self-schema, 120,129
group-attitude schema, 62-3
group-schema, 66-7, 314
groups, 29-30,141-6
co-ordinates of social, 70
and ideologies, 158-60
role in development of ideologies, 135,
140-60
versus members, 147-51
habitus, 47
Hall, Stuart, definition of ideology, 9
hegemony
Gramsci's, 3, 15, 102, 140
ideological, 260-1, 274-5
Herrnstein, R J., 297-8
historical knowledge, 31-2
history of science, 108
hyperbole, 273, 301, 303
hypothesis testing, 86
ideas, 15-16
and beliefs, 15-27
everyday meanings, 15-16
systems, 5, 26
identification
and social structure and organization, 124
use of term, 121-2
identities, multiple, 72, 85-6
and conflicting ideologies, 150-1
identity, 118-25
defining, 118-19
see also group identity; personal identity;
social identity
ideological discourse structures, 263-76
context constraints, 264-6
interaction strategies, 273-4
local meaning, 266-71
manipulation, 274-5
rhetoric, 272-3
369
schemata, 271
style, 271-2
topics, 266
ideological enemy, 278-9, 306-9
ideological groups, 14 -5
ideological influence, 245-54, 318
cognitive conditions, 246-7
generalization and abstraction, 252-4
opinion discourse understanding, 248-52
social conditions, 247-8
ideological institutions see institutions,
ideological
ideological leaming, 91,180-1
ideological reproduction see reproduction
'ideological squaré , 267, 301
idéologie, 1-2
ideologies
and attitudes compared structurally, 65-7
as axiomatic basis of shared social
representations, 126-8, 314
as beliefs, 26-7
cognitive definition, 48-52
as foundation of group beliefs, 48, 49-52
as general social beliefs, 32-3
and groups, 158-60
indirect links to discourse, 85-8
presuppose specificity for group or culture,
40
reconstructing, 284-90
relation to 'common sense', 106-7
social analysis, 315-16
socio-cognitive analysis, 313-15
typology, 69-70
and values, 69, 76-7
without groups, 154-9
ideologues, 16, 92, 99, 106, 149, 172, 174,
277
ideology
commonsense conceptions, 2
and discourse, 132, 191-9, 316-18
and discourse control, 210
fuzzy life, 1
as group identity, 120-2
Hall's definition, 9
and legitimation, 257-60
multidisciplinary theory, vii-viii, 4-7, 136,
193, 313
limitations and prospecta, 318-20
new concept, 8-9
primacy over action, 164-8
producing, 239-41
and society, 135-9
tradicional approaches, 1-4
versus knowledge, 108-9
ideology formation, social constraints on, 51
370
Index
implication, 207, 269
implicitness
and explicitness, 207
versus explicitness, 268-9
imposition, 180-1
inclusion, and exclusion, 159-60, 161
inculcation see ideological leaming
individual differences, 30, 93,155-6
individualism, 10, 131
inference, 84, 192
influence, different types, 244-5
information
expression or suppression, 267
level of description, 267-8
information and communication societies,
162-3
information processing, 17, 21
conscious and automatic, 100
metaphor, 47-8, 235
institutional contexts, legitimation and, 255-6
institutional racism. 189-90
institutionalization, degree of, 146, 316
institutions, ideological, 92, 135, 145, 186-90
Intelligenz,freischwebende, 3
intension, 204
intention, 216-18
interaction, 6, 209
and type of speech event, 215
interaction analysis, 206
interaction strategies, in ideological discourse
structures, 273-4
interaction values, 75
interactionist theories, 10, 18, 131
interests, power and, 8
intergroup discourse, 125
interiorization, 260
interpretation of discourse, 79-80
intersubjectivity, 36
intertextual complexes, 219-20
intragroup discourse, 125
'invented' ideologies, 98, 135, 172-4
IQ debate, D'Souza on, 297-9
irony, 272
judgements
ideologies control group specific, 116-17
and opinions, 19-20
justification, 109-10
King, Martin Luther, 295, 299-300
knowledge
and beliefs,18-19, 315
defined, 109-10
explicit and implicit, 99
nature of, 25, 109-10
and opinions, 33-6,111-14
and power, 114-16
relations between ideology and, 8, 108-9
social and political dimensions, 3
and truth, 108-17
language, relations with thought, 22-4,27
language analogy, 30, 32
language use see discourse
langue, 194
legitimation, 5, 8,255-62
defined. 255-7
delegitimation and discourse, 260-2
and ideology, 257-60
as political, 256-7
positive and negative, 11
of power abuse, 163, 165-8
values used as basis, 76-7
legitimization, of racism, 294-5, 318
levels
of abstraction, 17, 19, 24, 54, 207
of operation of ideologies, 92-3
structures and strategies, 200-1
lexical analysis, 205
lexical style, 205
lexicalization, 270-1
lexicon, 205
linguistics, 130, 194, 198, 200, 204
role of grammar in, 90
lived experiences', ideologies as, 103
Living with Racisin (Feagin and Sikes), 307
local meaning, in ideological discourse
structures, 266-71
Lukács, Georg, 2
macro-micro problem, 9, 136, 148-9, 191,
228-9
macrosociology, 186
manipulation, 5, 87, 97,183, 208, 217,
218-19, 258, 260-1
in ideological discourse structures, 274-5
Mannheim, K, 2, 182
Marx, Karl, 2, 96, 108, 179
Marxism, 140
mass communication research, 243
mass media see media
masses, 99
material resources, 143
materialism, 131
meaning
and interpretation, 204-5
knowledge and truth, 113-14
meaning of, 204
meanings, 203-7
as type of belief, 22
Index
media, 187-9, 262,265-6
and elite ideologies, 180-1, 184-5
members, versus groups, 147-51
membership, 224-5
and group categories, 151-60
membership ideology, 69,70
memory
beliefs and, 21, 205
see also episodic memory; social memory
mental dimensions of ideologies, 26, 27, 313
'mental frameworks' (Hall), 9
mental models, 45, 79-82,133, 205, 212-14,
249-50, 264, 315
control of in racism, 299-300
interface with social practices, 79, 129,
319
'mental' objects, 17, 18
mentalism, 10, 18, 131
meta-ideology, 145, 284, 318
metaphors, 272
micro level see macro-micro problem
microsociology, 53, 102, 186, 199
mind
influencing, 243-4
versus body, 17-18
mind control, 162-3, 273-5
minds, 16-18, 235
mitigation, 270, 273, 294, 309-10
models, 79-80,133
selectiva activation of old, 86, 87
moral order of society, 34
moral principies, 40
morphology, 202
Moscovici, Serge, 105
multiculturalism, 296-7
Murray, C., 297-8
neo-Marxism, 3
networks, neural, 23
neuroscience, 17
news values, 187-8
nodes, beliefs as collection of, 23
nominalization, 270-1
non-verbal discourse, 197-8
normalization, 91, 252
norms, 256, 257, 259
objective knowledge, and subjective
knowledge, 41-2
opinion
differences, 42
or knowledge, 111-14
opinion discourse understanding, 248-52
opinion management, 274-5
opinions, 39-40, 245, 250-1, 253, 315
and context models, 212
371
and judgements, 19-20
and knowledge, 33-6
personal, 33, 44-5
social, 33
and use of models, 85-6
oppositional ideologies, 11, 98, 100-1,
168-70
order of discourse, 196
organization of ideologies, 65-73, 316
organizations see institutions
piel processing, 57-8
parole, 194
participant roles, 220-3, 225-6, 264-5,
273-4
particular beliefs, versus general beliefs, 31-2,
41
'pathologies' of black culture, 277-312
performance, 194
personal, Iinking social and, 78-9, 83-5,
86-9
personal beliefs, versus social beliefs, 28-33,
314, 412
personal identity
forms, 119-20
and group identity, 118,119-25
personal memory see episodic memory
persuasion, 243-54, 258, 260, 263-4, 318
different types, 244-5
persuasive power, 162-3, 166-8
philosophy, 2, 5,130, 313
plans, 80
Plato,19
points of view see opinions
polarization, in-groups and out-groups, 161
policies, translation of ideologies, 189-90,
312
political cognition, 28, 132, 313
political parties, racist, 189
political psychology, 172
politics, 5,262,265-6
position ideology, 70,161
post hoc justification, 166
postmodernism, 197
power, 2,258
and domination, 161-4
and knowledge, 114-16
power abuse, 11, 163, 165-8, 262,273-4
practical competence, versus abstraer
competence, 54-5
pragmatics, 198, 200, 206, 211, 212
and context models, 83
'preferred' interpretations, 250-1
'preferred' mental models, 260-1, 264,
267-9
372
prejudice, 44, 61, 266, 291, 303
presupposition, 30-1,102-3, 207, 269
presupposition test, 39-40
problem/solution structure, 66, 67
professional groups, 145, 152-3
professional ideology, 70-1, 161
professional knowledge, 50
professional role, 221-2
'projection rules', 25
propaganda, 87, 99,184
propositions, 22-3, 206
protest demonstrations, 142
prototype theory, 155
psycholinguistics, 54
psychologism, 10
psychology, 4, 6, 29, 53, 131-2, 200
public discourse, 265-6
access, 173-4, 261-2
purpose, 218
race relations, USA, 277-312, 318
racism
ideology and discourse, 277-312
reproduction by discourse, 11-12
social dimension, 138-9, 157
see also ethnocentrism/racism
racist discourse, 300-11
racist groups, 156-60
racist ideologies, 11-12, 138
Rainbow Coaütion, 182
reconstructing ideologies, 284-90
reductionism, 17, 27, 43-4
referente, 204
referents, discourse, 225
reification, 17
relationships, importante, 4, 5-6
relativism, 36, 37, 40-1
ideological, 109, 110-14
strong forro, 110-11
weak form, 111
see also cultural relativism
relativity principie, 51
relevance entena, 130, 212, 214
representations
negative of others, 287-8, 317
social and mental 9-10
see also mental models; social
representations
reproduction, 8, 163, 228-34
defined, 228-30
and discourse, 230-4
organizing, 186-7
principies, 267
role of discourse, 191-3
Index
resistance. 11, 68, 70, 72,130, 168-70, 258,
316
resource ideology, 70
rhetorical structures, 208, 272-3
rhyme, 272, 273
ritual s, 26
roles, 220-3, 227
ruling -class ideology, 2, 15, 96,140,179,
255
schemata, 56-8, 133
schematic structures of discourse, 207-8
Schutz, Alfred, 102
science, ideology, 3
scientific knowledge, 106
scope, 265-6
scripts, 58-9,133
self
as abstract personal knowledge, 32
and context models, 82, 214
representations. 118
self-presentation, positive, 267, 301, 309-11,
317
self-servingness of ideologies, 8, 15, 68-9,
76
semantic macrostructures see topics
semantic memory, 29, 314
semantics, 203-7
and event models, 83
seroiotic terminology, 197-8
sequening, control, 273-4
shared ideas, 15-16
shared nature of ideologies, 29-30
sharing, 149-50
siga, 197
Sikes, Melvin, 307
situation modele see event models
s ial , 164-5
ooccial
ocial
and cognitive, 9-10, 51-2, 135-7
linking with personal, 78-9, 83-5, 86-9
social beliefs, 8,28-52,29-31
presupposed, 30-1
versus personal beliefs, 28-33, 41, 314
social categories, versus social groups, 144-7
social cognition, 6, 47-8, 126-34
and general cultural knowledge, 39
relevance. 126
structures and strategies, 55-64
use of term, 47-8
social competence, 247
social confiict, 145
social dimensions of ideologies, 26, 313
social domains, text and talk, 196-7
social enemy, 279-81
Index
social forrnations,140
social functions of ideologies, 69, 137-8,
316
social groups
characteristics, 154
versus social categories, 144-7
social identity, 141
as collective 'feeling', 122
contextually variable manifestations,
124-5
defined, 121-2
other ways of defining, 123-5
social knowledge, propositions with variables,
31
social memory, 29-31
social movements, 144, 153, 155, 174, 259
social position, 153
social practices, 6
and ideologies, 9, 78-9, 83-5, 316
interface with mental models, 129
social psychology, 47, 59, 61-4, 105, 132,
243
social relations, 6
social representations, 6, 8, 46-7, 133, 313,
314
and 'common sense', 105
and context, 225-7
personal models, 129
ideologies as, 127-8
shared, 141-4
use of term, 97
social role, 222-3
social structures, 6
socialization, 148-9, 193, 246
society, 135-90
discourse, and cognition triangle. vii, 5, 11,
131,136-7, 313
and ideology, 135-9
sociocultural beliefs see social beliefs
sociolinguistics, 54, 198, 211, 212
sociological facts )urkheim),108
sociology, 2, 5, 6, 47, 130
sound, 202
specific beliefs see particular beliefs
speech acts, 208-9, 218
stereotypes, 266, 302
story organization, 66
strategic analysis, 53-4
strategies, and structures, 53-64, 200-1
strategies of ideological control, 183, 184-5
structural analysis, 53-4, 56
'structural' learning, 245
structuralist approach, 53-64, 55
structures, and strategies, 53-64, 200-1
(
373
structures of ideologies, 55, 65-73, 128-9
categories, 69-70
ethnocentrism/racism, 288-90
struggle, and conflict, 145, 168-70
style, in ideological discourse structures,
271-2
stylistics, 202-3
subcultures, 39
subjective knowledge, and objective
knowledge, 41-2
surface structures, 201-4, 272
survival, 11
symbolic power, 262
'symbolic racism, 278
symbolic resources, 143
symbols, 26
syntax, 202-3
systems of ideas, 5, 26
taken-for-grantedness. 102-3
talk see text and talk
text and talk, 6, 319-20
global form, 207-8
mental representations and, 236-9
props and relevant objects, 220
of social domains, 196-7
see also discourse
theoretical framework, 4-7
open problems, 132-4
relevance, 130
theory of ideology, characteristics, 4-5, 130
thought, relations with language, 22-4, 27
tokens, and types, 84,194-6, 229
top-down legitimation, 174-6, 257
topic management, 273
topics, 206,207,237-8,266
triangle: cognition-society-discourse, vii, 5, 11,
131,136-7, 313
ttuth
and knowledge, 108-17
social and political dimensions, 3
truth criterio, 19, 25, 315
control over, 262
and knowledge, 34,109-10
and opinions, 33, 49-50
and social context, 36
versus evaluation criterio, 34-6, 41
truth values, as propositional functions, 42
truth/falsity debate, 130, 250
tumtaking, control, 273-4
types, and tokens, 84,194-6, 229
'underclass', 284, 302-3
Us and Them, 2, 68-9. 116-17,129,159-60,
247-8, 257, 267, 275-6
in racism, 278-82, 289-90, 312
value systems, 74-6
values, 69, 74-7, 256, 257, 259
with different ideological content, 76-7
historical, 76
and ideologies, 76-7
opinions and, 34
see also evaluation criteria
versus consistency, 90
verbal products, versus communicative events
193-4
see also non-verbal discourse
West, Comen, 308
Westem civilization, portrayal, 310-11
written text, 200
Notes
Chapter 1
1 Indeed, few scholars today would daim to practise a 'science of ideas', although there
are some who come close, such as French sociologist (pbilosopher, etc.) Edgar Morin, whose
four-volume sequence La méthode, ends with a book on Les idées: leur habitat, leur vie, leur
moeurs, leur organisation, in which also the 'organization of ideas' (the object of the
discipline of 'noology) is studied (Morin, 1991). Of course, there are historical antecedents
here, for example, in phenomenology, such as Husserl's book Ideas (Husserl, 1962).
2 Indeed, most studies of ideology in philosophy and the social sciences have a prominent
historical dimension. This is less the case for work on ideology in psychology, anthropology
and linguistica, which in general are less historically oriented. Since many of these studies will
be referred to more specifically in the next chapters, we here only mention the most prominent
books that provide such a historical background: Abercrombie et al. (1980,1990); Billig
(1982); CCCS (1978); Eagleton (1991); Freeden (1996); Kinloch (1981); Larrain (1979);
Manning (1980); Meszaros (1989); Rosenberg (1988); Rossi-Landi (1978); Seliger (1976,
1979); Skidmore (1993); Thompson (1984, 1990); Zcitlin (1994).
3 For a review of this 'restrictive' concept of ideology, see especially Seliger (1979), who
critically discurses the work of Bel (1960), Lipset (1960, 1972), Sartori (1966, 1969) and
Shils (1958), among others. See also the critical comments of Geertz (1973) on the pejorative
use of the concept of ideology.
4 See Marx and Engels (1974).
5 For a discussion of these contemporary changes in the theory of the relations between
superstructure and infrastructure, see, for example Wuthnow (1992).
6 A particularly interesting collection of studies documenting this evolution of European
neo-Marxism within British cultural studies, and especially within the work of Stuart Hall,
may be found in Morley and Chen (1996).
7 One major study that advocates such a more inclusive concept of ideology, and one of
the few systematically theoretical approaches to ideology, is that of Seliger (1979), who
defines ideology as a 'group of beliefs and disbeliefs expressed in value sentences, appeal
sentences and explanatory statements'. These sentences may refer to moral and technical
norms, and express views that relate to human relationships and socio-political organization.
Such an ideology may legitimate 'concerted action for the preservation, reform, destruction or
reconstruction of a given order' (Seliger, 1979: 119-20). Many of these contemporary debates
have their roots in the detailed theoretical analysis of Karl Mannhcim's Ideology and Utopia
(1936), which also discurses the distinction between evaluative and non-evaluative ideologies.
Also Mannheim thus emphasizes the role of ideologies in the context of the 'collective action'
of diversely organized groups.
8 For discussion on such political belief systems, see Chapter 2, and Chapter 2, Note 8, for
references.
9 See, for example, Rosenberg (1988) for such a psychological (Piagetian) approach to
ideology. See also the references in the next chapters.
10 Of course, as also Geertz (1973) points out, ideologies are not always rooted in, or
devised in order to legitimate, interests and power. They may also be a response to social
problems and contradictions ('strains') as lived and experienced by social members. At the
same time, the analysis in Chis book responds to a critical condusion of Geertz that both
322
Notes
approaches are indequate while failing to formulate in detail how the trick is really done', that
is, how exactly interests are related to ideology, and how social contradictions are 'symbol
ically expressed': 'Both interest theory and strain theory go directly from source analysis to
consequence analysis without ever seriously e xamini g ideologies as systems of interacting
symbols, as patterns of interworking meanings' (Geertz,1973: 207).
11 One example of a recent text in which ideologies (of contemporary movements) are
simply defined as 'discourses' is Garner (1996: 15).
12 Note that 'social cognition' in this book is not used (only) in the restrictive sense of the
information-processing approach, prevalent especially in the USA to the study of the social
mind (for survey, see, e.g., Fiske and Taylor, 1991), in opposition to the various European
approaches in social psychology, for instance on social identity, social categorization or social
representations (see, e.g., Fan and Moscovici, 1984; Tajfel, 1981; Spears et al., 1997). Rather,
I advocate an integration of these two approaches. For discussion and further references see
the chapters in Part I.
13 Such ideologies of opposition or resistance may of course be given a different name.
Thus, for instance Mannheim (1936) distinguished ideologies and utopias, the latter being
belief systems 'for a better world' which we also will call ideologies.
14 This position, currently formulated especially within the framework of Critical Discourse Analysis (CDA), has been explained in more detail in van Dijk (1993b). See also
Fairclough (1995), Wodak (1989, 1996).
15 For a discussion of the curcent relevance of Critical Theory and its relations to ideology,
see, for example, Agger (1991, 1992); Bailey (1994); Rasmussen (1996). For a discussion of
this critical study of ideologies and social inequality in the 'postmodern' world, see Simons
and Billig (1994). See also Larrain (1994) and Morley and Chen (1996) for a discussion of the
postmodern critique of ideology. See Ibáñez and Iñiguez (1997) for a collection of work in
critical social psychology. Note though that the term'critical' in these various studies has
rather divergent meanings and applications.
16 This does not mean, of course, that there is no carlier work on racist ideologies, but only
that there is as yet no work on racist ideologies that uses the framework proposed here. See,
for example, Note 6 of Chapter 19, for scientific ideologies of race. For the relations between
discourse and racist ideologies, see, for example, van Dijk (1984, 1987); Wodak et al. (1990),
and references in various later chapters.
Chapter 2
1 One of the (vast) arcas of research that will be largely ignored in this book is that of the
listory of ideas', and related fic1ds of historical inquiry, such as the study of 'mentalities'.
See, for example, Lerner (1991).
2 The mind-body debate keeps haunting cognitive science, if only as a pseudo-problem.
For recent discussion, see, for example, Warner and Szubka (1994). Interestingly, most
psychologists simply ignore the question about the 'existence' of the mind, and go about their
everyday business of describing and explaining psychological phenomena with the tacit
assumption that minds do exist. The ongoing debate, especially among philosophers and
neuroscientists mainly involves the relations between mind and brain. See, among many other
contemporary studies: Clancey et al. (1994); Kosslyn and Koenig (1992); Pinker (1994);
Searle (1992, 1995).
3 For an explicit social-developmental (Piagetian) approach to ideology, see Rosenberg
(1988).
4 Such'interactionism' may be found in certain directions of research in ethnomethodology and discursive psychology, in which the (socially) 'real' things we need to deal with are
interaction and discourse. That is, whatever'mental' things may exist, they are relevant only
through their expression or formulation in social practices, text and talle (Coulter, 1979, 1983,
1989; Edwards, 1997; Edwards and Potter, 1992; Harré, 1995; Harré and Stearns,1995; Potter
and Wetherell,1987). See also the critique of (mental) 'representationalism' in Shanon (1993).
Notes
323
These directions of research deserve detailed analysis, which is, however, beyond the scope of
this book. I hope to return to this issue in a future publication. As is shown in this section of
the book, I recognize the relevance of a proper cognitive analysis, but agree with the discursive
psychologists that discourse and interaction play a fundamental role in the acquisition and the
structures of 'mental' phenomena, such as knowledge and ideologies. Also, I fully agree with
them that most of traditional social psychology and of course most cognitive psychology has
ignored the relevance of the influence of social structures in psychological studies.
5 The nature of beliefs and their relation to knowledge continues to be discussed mainly in
epistemology, and much less in psychology itself, as is often the case for such fundamental
notions. See, for example, Kornblith (1994); Lehrer (1990). We shall later come back to the
discussion of belief systems in social and political psychology.
6 For a review of discussions on the nature of emotion, see, for example, Frijda (1987);
Ortony et al. (1988).
7 For classical and recent studies that define our current thinking on mind and memory,
see (among many other books), Ashcraft (1994); Barsalou (1992); Cohen et al. (1993); Kintsch
(1977); Neisser (1982); Solomon et al. (1989); Tulving (1983).
8 Many cognitive approaches to memory and beliefs (see Notes 5 and 7) assume such a
network representation, even when they (also) use propositions for practical purposes of
description. A more recent, neurologically inspired reformulation of the network idea may be
found in connectionist cognitive psychology (McClelland and Rumelhart, 1986; Rumelhart
and McClelland, 1986). Here the linear processing metaphor of classical computers is replaced
by the parallel processing metaphor of 'neuro-computers'.
9 The question of basic beliefs is here related to that of 'basic acts', which also has been
discussed in psychology. See Newtson (1973). For a philosophical analysis of basic actions
(using an example similar to the one we used — about the Balkans), see Searle, 1983:
99-100).
10 The nature and structure of such ideological (and other) belief systems have been
studied by, for example, Abelson (1973); Carlton (1984); Converse (1964); Little and Smith
(1988); Tetlock (1984, 1989); Wegman (1981). For the difference between such approaches
and cognitive psychological approaches to beliefs, see Quackenbush (1989).
11 Among the theorists of ideology who reject mere study of belief systems and who plead
for a combination with social, interactional, dimensions, is Rosenberg (1988).
12 See., for example, Geertz (1973); Oberschall (1993); Wutbnow (1989).
13 For a discussion of this debate (about the linguistic relativity thesis), see, for example,
Lucy (1992).
Chapter 3
1 A classical study of political belief systems is Converse (1964). For current discussions
and critique, see, for example, Iyengar and McGuire (1993); Lau and Sears (1986).
2 The limerick opens Salman Rushdie's book The Jaguar Smile: A Nicaraguan Journey,
London: Pan, 1987.
3 See Tulving (1983) for a classical theory of episodio memory and its distinction from
semantic' memory.
4 For the 'cleeper' neuroscientific approaches to the architecture of the mind and its
various constructs, see the notes in the previous chapter.
5 As we shall see in more detall later, part of social memory, namely, knowledge, tends to
be studied by cognitive psychologists, whereas other socially shared beliefs (such as attitudes)
are the domain of study of social psychology. Social memory is currently studied especially in
cognitive social psychology (Devine et al., 1994; Fiske and Taylor, 1991; Forgas, 1981;
Resnick et al., 1991). Although most studies of social cognition in the USA are oriented
towards cognitive psychology (an 'individualistic bias' criticized by more socially oriented
social psychologists), this love is unfortunately hardly returned by the cognitive psychologists,
who generally ignore the many forms of lot (social) cognition'. This situation is among the
324
Notes
many explanatory factors why much of contemporary psychology (and the same is trae for
cognitive science) is socially and culturally rather underdeveloped. It is especially under the
inspiration of Soviet psychology that the socio-cultural study of cognition has been able to find
a small, but important, niche in Western (and especially US) psychology (see, e.g., Hickmann,
1987; Wertsch, 1985; Wertsch et al., 1994).
6 Within the framework of his sociology of knowledge, Mannheim (1936: 2) already
emphasized that ideologies cannot be explained in terms of personal beliefs, but have a social
nature. In his argument he also uses language for comparison to show that although language
may be used individually as 'speech', people use a language as a system that is socially and
historically shaped. At the same time, Mannheim wams that the notion of social thought does
not imply that there is something like a 'group mind'. Thus, in his words, ideology is the 'style
of thought' of (the members) of a group (p. 3). He distinguishes between particular (personal,
individual) ideologies, for example as distorted views of reality arising from people's life
situation, on the one hand, and inclusive, total conceptions of ideology, which are the
ideologies of an age or a group. The latter are widely diverging thought systems, which give
rise to totally different modes of experience and interpretation (p. 51).
7 One of the authors writing on ideology who emphasize that ideologies are a group
characteristic is Scarbrough (1990).
8 See, for example, the contributions in Lau and Sears (1986) for discussion about
whether or not ideologies actually exist as 'belief systems'.
9 The notion of 'factual belief is of course a theoretical one, not a commonsense notion,
given the fact that in everyday language use 'beliefs' are associated with doubtful knowledge
or (mere) opinions, so that 'factual beliefs' would be a contradiction. We use the notion in
order to emphasize the general notion of 'belief, and in order to be able to differentiate
between different kinds of belief. See also the discussion in Chapter 11.
10 For various approaches to the discourse marking of evidentiality, perspectiva and
opinion, see, for example, Biber and Finegan (1989); Mayar, (1990); Schieffelin (1996).
11 This is the approach advocated by discursive psychology. See, for example, Edwards
(1996) for detail.
12 For a detailed discussion of che tradicional notion of attitude, see Eagly and Chaiken
(1993). Note that most traditional approaches to attitude make no clear distinction between
social and personal opinions, or specific and general opinions. Jaspars and Fraser (1984)
criticized che individualistic approach to attitudes in much of social psychology, and uscfu ly
remind us of che fact that che original notion of attitude refers to socially shared beliefs of a
group. This will also be my approach, thereby adding that such social attitudes (e.g. about
abortion or nuclear energy) are not isolated beliefs, but complex structures We shall come
back to che notion of 'attitude' and its structures in che next chapter.
13 There is now a considerable literature in this kind of discursive, rh torical and social
constructionist psychology. For some key texts in which diese claims are formulated, see, for
example, Billig (1987, 1991b, 1995b); Billig et al. (1988); Edwards (1997); Edwards and
Potter (1992); Harré (1995); Harré and Gillett (1994); Potter (1996); Potter and Wetherell
(1987).
14 One theory of 'social representations' is usually associated with che work of Serge
Moscovici (Paris) and his followers. See, for instance, Augoustinos and Walker (1995);
Breakwell and Canter (1993); Farr and Moscovici (1984). For an account of ideology in terms
of social representations, see Aebischer et al. (1992); Augoustinos and Walker (1995). The
French theory of social representations, however, is more specific than our general use of che
term (as socially shared beliefs), and especially applies to mundane, commonsense uses of
scientific knowledge in everyday life, for instance che lay uses of psychoanalysis. There has
also been considerable critique of che notion of social representation. See, for example, Jahoda
(1988) and che reply by Moscovici (1988).
15 The notion of habitus was introduced by French sociologist Pierre Bourdicu. See, among
many of his writings, for example, Bourdicu (1985, 1988, 1990). For (sociological) critique,
see, for example, Alexander (1995).
Notes
325
16 For another example of a more integrated approach to social cognition, see Augostinos
and Walker (1995). For US approaches to social cognition, see, for example, Devine et al.
(1994); Fiske and Taylor (1991); Higgins et al. (1981); Wyer and Srull (1984, 1989). For
(mostly European) work on social identity, social categorization, social represenations, and
intergroup relations, see, for example, Farr and Moscovici (1984); Forgas (1981); Spears et al.
(1997); Tajfel (1978, 1981); Turner and GIles (1981). See also the references in Note 4 and the
references in the next chapter.
17 There are some (not very detailed) suggestions in the literature that take ideologies as
the general organizational basis for attitudes, and that define attitudes as more-specific
opinions about issues or social domains. See, for instance, Scarbrough (1984, 1990), who also
discusses the relations between ideologies, attitudes and social representations.
18 See especially the relevant work of Foucault about these relationships between (medical)
knowledge and power: for example, Foucault (1975, 1980).
Chapter 4
1 As far as I know, diere are no detailed general studies of the difference between the
structural' and the'dynamic' approach in the humanities and the social sciences. For a
discussion of the distinction in psychology, see van Dijk and Kintsch (1983). In sociology, this
debate typically sets apart microsociological and ethnomethodological approaches from earler
'structural' or 'functional' sociology. See, among many other studies, Button (1991) and
Heritage (1987).
2 Schema-theory in cognitive science essentially goes back to Barden (1932), who
assumed that knowledge is represented in a schematic fashion. Its most influential formulation
in contemporary psychology has been in tercos of knowledge 'scripts as introduced by Schank
and Abelson (1977), after earlier notions such as that of scenarios' (Charniak, 1972) and
'frames' (Bobrow and Collins, 1975).
3 For discussion, see the recent debate on connectionist, neural and parallel processing:
Baumgartner and Payr (1995), Clancey et al. (1994), Rumelhart and McClelland (1986). This
work only very recently begins to influence social psychology and the theory of attitudes and
social representations (Eiser, 1994).
4 See Schank and Abelson (1977).
5 Such smaller units, sometimes called Memory Organization Packages (MOPS), are for
instance discussed in Schank (1982).
6 This is especially the case in the 'social cognition' approach in the USA. See, for
example, Fiske and Taylor (1991). For a discussion of the early uses of 'schemata' in social
cognition, see Brewer and Nakamura (1984), Higgins et al. (1981). For the use of schemata in
political psychology, see Kuklinslci et al. (1991).
7 Representations for opinions and attitudes have especially been attempted by Robert
Abelson, who may also be credited with the invention of the script-concept (see, e.g., Abelson,
1973, 1976, 1981; and bis very early attempt with Rosenberg in Abelson and Rosenberg,
1958). For a survey of other attempts to model attitudes and other evaluative social cognition,
see Fiske and Taylor (1991) and Eagly and Chaiken 1993).
8 These natural, but fallible, forms of information processing and social judgement have
been extensively studied in psychology. See, for example, Arkes and Hammond (1986);
Kahneman et al. (1982); Martin and Tesser (1992); Nisbett and Ross (1980).
9 Although this three-component distinction may be found in many studies of attitudes
(for survey, see Eagly and Chaiken, 1993), it was in fact seldom empirically tested (see,
however, Breckler, 1984).
10 For classical studies in social psychology on the consistency, balance or cognitive
dissonance of opinions, see, for example, Abelson et al. (1968); Heider (1946, 1958); Festinger
(1957). More recent approaches are collected in Pratkanis et al. (1989). For a recent survey of
classical as well as modem approaches, see Eagly and Chaiken (1993).
326
Notes
Chapter 5
1 As noted several times before, the vast literature on ideologies has seldom been
concerned with the detailed intemal organization of ideologies. Ideology descriptions, when
given at all, tend to be largely impressionistic from this structural point of view, that is,
summaries or stories about the beliefs of groups. Since also opinions, attitudes and ideologies
are not always distinguished, some proposals for ideological 'structures' are in fact proposals
for attitude structures. For some more explicit attempts that speculate about ideological
organization, see, for example, Seliger (1979), who organizes ideologies in tercos of the nature
of the kinds of statements they contain: a circle of descriptions, analysis, implementation and
rejections, with moral or technical prescriptions in the centre. See also Roseman (1994).
2 See Abelson (1976) for a script-like approach to attitudes.
3 For a detailed analysis of such opinions about immigration and immigrants, see van Dijk
(1984, 1987, 1991, 1993a).
4 The organization of (panty) ideologies as a function of the perceived (importance of)
problems has been studied by Van Schuur (1984).
5 These narrative structure categories have first been introduced by Labov and Waletzky
(1967) and Labov (1972). Later it was also applied in the analysis of our knowledge about
such narrative structures, and indeed in the study of story comprehension, although also other
structures (for instance in terms of actions, events and goals) have been proposed (see, e.g.,
Mandler, 1984; van Dijk, 1980).
6 Despite the occasional reference in social cognition research to the notion of 'group
schemá , no such schemata have ever been described in any detall, as far as I know. There are
some research suggestions that groups are represented much like persons, but no structural
description is given in that case (Wyer and Gordon, 1984; Fiske and Taylor, 1991: 327). The
present schema is derived from my earlier work on the structure of ethnic attitudes on
minorities (van Dijk, 1984, 1987).
7 See, however, Billig' s work on 'rhetorical thinking, in which thought itself is assumed to
be organized in a rhetorical or argumentative way (Billig, 1987, 1991b).
8 One empirical study of two possibly conflicting ideologies is that of Eckhardt et al.
(1992), which shows that religious scholars seem to be able to handle coexisting religious and
scientific belief systems without much personal conflict.
Chapter 6
1 The literature on values is vast. As is the case for other social representations, however,
we know as yet very little about their precise cognitive nature. How, indeed, are they
represented in social memory? Probably not just as the word-concepts (such as Ilappiness' or
'Justice') used here as well as in other studies. They may be complex mental representations
(indeed, the whole complex'idea of Justice') that are merely conveniently 'summarized' by
such concept-words, so that they can be easiy expressed in communication and interaction.
For examples of classical and modem studies, see Brewster Smith (1969); Eisenberg et al.
(1989); Hechter et al. (1993); Hofstede (1980); Rokeach (1973, 1979). Schwartz and Bilsky
(1990) propose an empirical theory of universal content and structure of human values defined
as people's conceptions of the goals that serve as guiding principles in their lives, that is, a
(small) number of universal motivational concems (such as hedonism, achievement, power,
security, etc.). Even when I assume that values are mentally represented, I would emphasize
the socio-cultural nature of such values instead of their individual (motivational) dimension.
Chapter 7
1 The theory of mental models has been developed in psychology, and especially also in the
theory of text comprehension, sine the early 1980s Qohnson-Laird, 1983; van Dijk and
Notes
327
Kintsch, 1983; van Oostendorp and Zwaan, 1994). Some of its more logical aspects go back to
abstract model theories that were designed as semantics for formal languages. Such formal
models (or model structures) feature, for example, the set of individuals referred to by the
expressions of formal statements. Mental models, although theoretically little developed,
should be more 'realistic' in the sense that they must account for the typical types of objects
that define possible situations.
2 So far, there is no theoretical work on experience models as such (see van Dijk, 1997).
However, most literature on episodic memory, and about'autobiographic' memory for events,
actions, persons, episodes and personal experiences, will provide some elements of such a
theory (see, e.g., Neisser and Fivush, 1994; Rubin, 1986; Srull and Wyer, 1993; Tulving,
1983). Our point here is first that the subjective representation of episodes should be framed in
terms of models and model structures. Second, we want to emphasize that such models also
play a role in the production and understanding of social practices in general and in discourse
in particular. We thus want to unify the usual accounts of 'situation models' in the textprocessing literature and the work on episodic representations of actions and events as well as
on autobiographical memory. In other words, also cognitively (and not only socially) the way
we engage in, and understand, text and talk should be made a part of a broader theory of our
everyday experiences.
3 Various sociological and psychological theories of social episodes provide suggestions
for the inclusion of basic categories into model schemata (see, e.g., Argyle et al., 1981; Forgas,
1979; Furnham and Argyle, 1981).
4 In linguistics, case grammars and other functional approaches to the study of the structure
of propositions and their syntactic expressions make use of such basic categories (Dik, 1978,
1989; Fillmore, 1968). At another level some of them also appear, as 'descriptions of
episodes', in theories of narrative (see Labov and Waletzky, 1967; Labov, 1972).
5 One of the studies that examines the representation of Self in relation to ideologies and
narrative structures, is Gregg (1991).
6 This seems to be the more intuitively'empiricar way of learning about the world.
However, there is evidence that 'learning from experience' through the generalization of
episodic models may not (always) be the way we acquire general knowledge. At least
fragments of semantic or social knowledge about the world may be acquired 'directly', that is
through discourse: for example, by the explanation of words or by generic sentences in stories,
arguments or other forros of discourse. At this point, the theory joins the more general theory
of knowledge acquisition and learning, a vast field of cognitive, developmental and educational psychology that obviously cannot be discussed here. My point is only to show how
personal knowledge (models) about events can be related to socially shared beliefs. Although
much is known about processes and conditions of learning and social knowledge acquisition,
our insight into the detailed representations involved is as yet rather fragmentary, as we have
seen before for the notion of scripts and related concepts. Since much knowledge acquisition
and learning takes place on the basic of discourse, much of the relevant literature on text
comprehension deals with the same processes. Similarly, also much work in artificial
intelligence, dealing with the simulation of knowledge representation and acquisition, is
relevant here. Thus, broadly speaking, after the more behaviouristic reduction of learning in
terms of conditioning and stimulus generalization, and 'social leaming' approaches based on
them, most contemporary approaches are clearly cognitive, and formulated in tercos of various
formats of memory representations for knowledge and beliefs. For various surveys and other
studies that use the framework proposed here, see Freedle and Carroll (1972); Glaser (1987);
Gonzalvo et al. (1994); Mand1 and Levin (1989); Schank (1982); Schank and Abelson (1977);
Strube and Wender (1993); van Dijk and Kintsch (1983).
7 Also in relation to what has been said in Note 5, there exists evidence that shows that
social representations may sometimes be stronger than personal experiences (e.g. of poverty)
as a reason to support (e.g. economic) policies (Lau and Sears, 1981). This suggests either that
public arguments are integrated into the model, and thereby influence personal experience,
action and discourse, or else that social representations (attitudes) may influence discourse
directly, especially when diese are cometunicated by credible elite groups. Of course,
328
Notes
responses to (survey) questions may also show effects of compliance and consensus with what
'everybody thinks', and hesitations to show economic hardship. In sum, models, opinions,
social representations and the ways these are expressed or not, mitigated or not, in various
discourses and contexts, fomi a very complex combination, which simplistic results of public
opinion surveys cannot possibly begin to make explicit.
8 There are many theories of attitude formation and change, as well as theories of
persuasion, that deal with these processes, but seldom in terms of models. This is also because
much work on opinions and attitudes does not differentiate between personal, contextual
opinions as represented in models, and general, socially shared attitudes. For survey of the
more traditional approaches to attitude formation, see Eagly and Chaiken (1993).
9 See van Dijk (1984,1987) for studies about such forms of biased recall of negative
events in racist storytelling and argumentation. For general social psychological work on
hypothesis testing and the self-confirmation of social stereotypes, see, e.g., Snyder (1981a,
1981b). Note though that this' phenomenon does not exclude that in specific contexts people
precisely tend to recall what is not consistent with their own group attitudes, e.g., when they
(better) recall the statements of their opponents, so as to be able to better refute these.
10 These assumptions may be found both in classical political science (Converse, 1964)
and current political cognition (Lau and Sears, 1986; Iyengar and McGuire, 1993), as well as
in discursive psychology (Billig et al., 1988).
Chapter 8
1 These assumptions may be found both in classical political science (Converse, 1964) and
current political cognition (Lau and Sears, 1986; Iyengar and McGuire, 1993), and presently
also, in different guises in discursive and rhetorical psychology (e.g. Billig et al., 1988; Billig,
1991a, 1991b; Potter and Wetherell, 1987). More specifically for ideologies, Seliger (1979)
adopts a middle position: He assumes that ideologies are structured, but not fully consistent.
Also Rosenberg (1988) emphasizes the structured nature of ideologies. In the literature on
political cognition, it is generally assumed that ideological consistency is a function of
expertise: Those who know more about politics simply have more consistent political attitudes,
and probably a more consistent underlying ideology (see, e.g., Judd and Downing, 1990).
Other research suggests that opinions expressed when an ideological schema is activated are
more coherent than when no such schema is activated, at least for people who are conscious
about their ideological orientation (Milburn, 1987).
2 Such questions have been raised at least since the classical theories of cognitive balance,
consistency and dissonance, mostly in social psychology, and often related to the study of
attitudes (Abelson, 1973, 1983; Abelson and Rosenberg, 1958; Abelson et al., 1968; Festinger,
1957; Heider, 1946, 1958; Rosenberg, 1960; Rosenberg et al., 1960). However, most of these
studies focused on relations between individual beliefs, and did not investigate the overall
structure of attitude systems and ideologies. Grofman and Hyman (1974) analyse the
systematicity of ideologies in terms of connectedness, consistency and coherence, and
conclude that, by these criterio, ideologies are indeed belief systems.
3 Among the many sociological and historical studies that pay attention to these institutional and organizational dimensions of ideology, see, for example, Douglas, 1986; Jones,
1984; Wuthnow, 1989.
4 For cognitive dissonance, see, e.g., the classical study by Festinger (1957). For a more
general discussion of this and other aspects of cognitive consistency, see Abelson et al.
(1968).
5 See my earlier empirical work on discourse and racism, based on data in several countries
and involving people from different socio-economic backgrounds and in different institutional
situations (van Dijk, 1984, 1987, 1991, 1993a).
6 For some studies of contemporary ideological change, see, for example, Adams, 1993;
Collins, 1992; Larana et al., 1994.
Notes
329
Chapter 9
1 Since this study avoids historical reviews of earlier conceptions of ideology, it is also
beyond its scope to examine the history of the notion of 'false consciousness'. For such a
historical study, see Lewy, 1982; Pines, 1993. See also Jost and Banaji, 1994; Wood, 1988.
See also the classical discussion in Mannheim (1936: 62ff.).
2 See the well-known debate about the 'dominant ideology hypothesis' (Abercrombie et al.,
1980,1990).
3 For studies of the example of class and class consciousness in relation to dominant
ideologies, see, for example, Giddens and Held, 1982; Joyce, 1995; Therborn, 1980.
4 Awareness has also been studied in political psychology, for example, in relation to the
issue of elites versus public opinion. See, for example, Zaller (1990).
5 For studies of group-consciousness (for example, of gender, class or ethnicity), see, for
example, Bell, 1995; Brooks, 1994; Davis and Robinson, 1991; Dillingham, 1981; Edwards,
1994; Graetz, 1986; Gurin and Townsend, 1986; Hall and Allen, 1989; King, 1988;
Lockwood, 1966; Rowbotham, 1973; Weakliem,1993.
6 See, for example, Lau and Sears, 1986.
7 For details, see, for example, Baars, 1988; Davies and Humphreys, 1993; Dennett, 1993;
Greenberg and Tobach, 1983; Jackendoff, 1987; Marcel and Bisiach, 1988.
8 For a study of ideological denial, see van Dijk, 1992.
Chapter 10
1 See Gramsci (1971). See also Adamson, 1983; Femia, 1987; Hall et al., 1978.
2 For a discussion of 'common sense' in ethnomethodology, see Eglin, 1979; Elliot, 1974;
Sharrock and Anderson, 1991. See also the other contributions in Button, 1991.
3 See, for example, Billig, 1991b; Billig and Sabucedo, 1994; Eagleton, 1991; Lewis,
1992.
4 On commoñ sense and social representations, see Billig and Sabucedo (1994); Purkhardt
(1993).
5 For the role of common sense in argumentation and accounts, see, for example, Antaki,
1994; Billig et al. (1988).
6 Furnham (1994) discusses these and several other terminological variations of the
concept of common sense, such as ' ordinary', 'lay' or 'folk' beliefs.
7 For the relations between common sense and scientific knowledge, see also Fletcher,
1993; Siegfried, 1994; Van Holthoon and Olson, 1987.
8 For a discussion of common sense in psychology, see Siegfried, 1994; Wegner and
Vallacher, 1981.
9 See, for example, Farr and Moscovici, 1984; Augoustinos and Walker, 1995.
10 For the role of common sense in the study of racism, see also Essed, 1987; Lawrence,
1982.
Chapter 11
1 For a discussion of the opposition between knowledge/science and ideology, see most
classical studies of ideology (see Note 1 of Chapter 1). See also Althusser, 1984; Aronowitz,
1988; Bailey, 1994; Larrain, 1979; Mannheim, 1936; Mészáros, 1989; Pines, 1993. See also
the next chapter.
2 See, for example, Button, 1991.
3 For these philosophical debates about knowledge and its foundations, see, for example,
Dancy, 1985; Kornblith,1994; Kruglanski, 1989; Lehrer,1990.
4 The study of 'opinions' stretches from philosophy (often in terms of 'beliefs' versus
Inowledge; see, e.g., Hintikka, 1962), to the study of public opinion in social psychology and
330
Notes
political science (e.g. Glasser and Salmon, 1995). Within his own 'rhetorical' framework,
Billig (1991b) discusses opinions in relation to ideology, and stresses the argumentative nature
of opinions instead of their cognitive properties. Billig (1989) also shows that there is a
difference between expressing weak or strong 'views', where people with strong views show
more variability.
5 On the relations of knowledge and power, and the truth-defining nature of institutions,
see, for example, Aronowitz (1988); Foucault (1972, 1980).
6 The relations between knowledge and ideologically based attitudes have been explored
also in social and political psychology. Thus, if feminist attitudes are important to women, also
their knowledge acquisition about gender relations may be deeper and more refined (see, e.g.,
Berent and Krosnick, 1995).
7 Mannheim (1936: 19) also argues that whether or not worldviews or ideologies may be
(objectively) 'false', they may serve to make 'coherene the fragmenta of reality as seen by the
group members that share such a worldview. Whether beliefs of a group are true or false, it is
their 'definition of the situation' that counts. Indeed, besides other criteria, group membership
for him implies that group members 'see the world' in terms of the meanings of the group. He
also emphasizes that for this same reason, more generally, knowledge is by definition related
to the viewpoint, position and interests of the group, and hence relative (or rather what he calls
'relationise) (p. 67ff.).
Chapter 12
1 The literature on social identity is vast, and cannot possibly be reviewed here. I limit my
discussion to social (group) identity and to the (close) relationship between such group identity
and ideology, that is to the question'Who are we?' (related to, but different from the question
about the social Self). For details, see, for example, Abrams and Hogg, 1990; Tajfel, 1981;
Tumer and Giles, 1981. It is remarkable, though, that Chis literature seldom speaks of
ideologies, and more generally it is not always clear in this social psychological literature
whether 'social identity is a property of individual social group members, ora sharedproperty
of a whole group. On the other hand, there is work on (new) social movements in which the
relations between ideology and identity are more clearly established. See, for example, the
studies in Larafia et al. (1994).
2 For the notion of 'self-schema', see Markus (1977).
3 On social identity, see, for example, Abrams and Hogg, 1990; Monis and Mueller,
1992.
4 See the discussion on the relations between social movements, social identity, ideology
and (mainly referrring to Mannheim's work) 'utopia' in Tumer (1994).
5 Melucci's approach to social identity and the self-definition of groups features similar
categories (see Melucci, 1996; and, for general discussion, Johnston et al., 1994).
6 This example of the 'transient identity' of the (Dutch) peace movement is discussed by
Klandermans (1994).
7 See Billig (1990) for how collective memory (about the Royal Family in the UK) may be
managed by ideologies, for example, what is remembered and what is ignored.
8 The relations between social groups and questions of social identity have especially been
emphasized for the 'new social movements' (NSMs) of the last decades, such as the peace
movement, the ethnic and women's movements, the gay rights movement, and various
nationalist movements. For diese NSMs the main reasons of their existence were not so muela
'structural', socio-economic (as was the case for the 'old' social movements, such as the
working-class movement), but especially also a question of identity, human rights or even
lifestyle, where individual and social claims may be merged. Especially also, questions of
symbolism and culture have been found to be characteristic of NSMs. Contrary to carlier
structuralist approaches to social movements, with their focus on socio-economic conditions
and opportunities, current analyses of the NSMs thus tend to emphasize the 'shared meanings'
involved in the self-definition of movements. For discussion, see, for example, Johnston et al.
Notes
331
(1994); McAdam (1994); Melucci (1989,1996). For a critical comment on the cognitive
shared frameworks' concept of Melucci, see Billig (1995b).
9 For a discussion of these discursive dimensions of the construction of social movements,
see, for example, Klandermans (1992). My own earlier work on discourse and racism shows
how racist and anti-racist groups and institutions are largely also constituted by text and talk
(van Dijk, 1984, 1987, 1991, 1993a).
Chapter 13
1 The critique that follows is not directed against individual writers, but to the various
tradicional and contemporary approaches to ideology we have been referring to in the past
chapters. Thus, my critique first addresses the various Marxist or neo-Marxist approaches,
which all but ignore psychological dimensions of ideologies (for surveys, see Eagleton, 1991;
Larrain, 1979; see also Fairclough, 1995). Second, it addresses some of the more radical tenets
of 'discursive psychology, which tend to reduce the mind to discourse (Edwards and Potter,
1992; Harré and Gillett, 1994; Harré and Stearns, 1995; Potter and Wetherell, 1987; Potter et
al. 1993; for ideology, see also a less radical position in Billig, 1991b).
2 For some of che critical positions summarized here, see, for example, Himmelweit and
Gaskell (1990); Resnick et al. (1991).
Chapter 14
1 See the references in Note 1 of Chapter 1.
2 Such a (unfortunately seldom heeded) plea for a cognitive sociology, has been eloquently
malle already by Cicourel (1973).
3 This conception of racism has been dealt with in more detall in van Dijk (1984, 1987,
1991, 1993). See also Essed (1991). Among the numerous other studies on racism that have
influenced my cónception are, for example, Barkan (1992); Barker (1981); Dovidio and
Gaertner (1986); Haghighat (1988); Katz and Taylor (1988); Miles (1989); Solomos (1993);
Solomos and Wrench (1993); Wellman (1993).
Chapter 15
1 As long ago as 1936, Mannheim (p. 3) emphasized that ideologies are shared by groups,
and function as a basis of collective action.
2 A useful survey of the classical and contemporary perspectives on class may be found in
Joyce (1995).
3 For these and other criteria of groups and group relations, see especially (mostly
European) social-psychological studies on inter-group theory, for example, Billig (1976);
Tajfel (1978, 1981); Turner and Giles (1981); Tumer et al. (1987).
4 For a similar analysis of the relationship between social identity and sharing representations, see, for example, Moscovici and Hewstone, 1983; Scarbrough, 1990.
5 Social movements and their ideologies are discussed by, for example, Laraña et al.
(1994), Melucci (1996) and Oberschall (1993). For the close relation between social identity,
ideology and new social movements, see the discussion and the references given in Chapter 12
(see also Note 6).
6 For corporate cultures, see, for example, Hofstede (1980). Business ideologies have been
studied by Goll and Zeitz (1991); Mattelart (1979); Mizruchi (1990); Neustadtl and Clawson
(1988); Rothman and Lichter (1984).
7 Indirectly and directly, diere is a massive literature on the relations between ideology and
social group conflict. Most studies mentioned earlier (see Chapter 1, Note 2), especially those
reviewing the Marxist tradition, of course emphasize the role of group (class) conflict as a
332
Notes
basis for ideologies. From a psychological perspective (about social dominance), see also
Sidanius (1993). More generally, see the following recent studies of social conflict, often also
discussed in relation to ideology: Feagin and Feagin (1994); Fisher (1990); Oberschall (1993);
Worchel and Simpson (1993). For the expression and enactment of conflict in discourse, see
Grimshaw (1990).
8 As with many other sociological issues dealt with in this and other chapters of this
section, we cannot go into the complex details of the macro—micro problem (often pseudoproblem) in the social sciences. What, among other issues, is at stake here are analyses of
different levels of social reality, which each may require its own theoretical framework, as is
the case for microsociological analyses of interaction and conversation on the one hand, and
larger social (group) relations and structures on the other hand, with several meso-levels in
between. That the problem, on one analysis, is often a pseudo-problem is because in concrete
analyses both macro- and micro-notions may be required. Thus, to study social power (such as
sexism or racism) in conversation, for instance, we obviously need a framework that combines
both levels. And as microsociologists often remind us, higher-level societal relations and
structures simply manifest themselves only at the level of everyday routines, practices, or
interactions of social actors. For further discussion, see, for example, Alexander et al. (1987);
Knorr-Cetina and Cicourel (1981); see also van Dijk (1980).
9 Relationships between group ideology and (often conflicting) identities have often been
studied, for example, by Garcia (1989); Gregg (1991); Hummon (1990); King (1991);
Lipiansky (1991); Oberschall (1993); Rees (1985); Rothstein (1991); Shotter and Gergen
(1989). For combined influences of race and class, see McDermott (1994).
10 This example of women and blacks in the media has been examined in several studies.
See, for example, Dines and Humez (1995); Milis (1988); Van Zoonen (1994); Wilson
(1991).
11 On class in general, see Joyce,1995. For an analysis of various forms of 'symbolic
capital' as constitutive of class, see especially the work by Pierre Bourdieu (e.g. Bourdieu,
1984a, 1989). McDermott (1994) shows how ideologies are often the result of the combined
influence of different group memberships, such as those of race and class.
12 Professional ideologies have been the object of much research, especially in the medical
and legal professions; see, for example, Byme (1993); Dickson (1993); Globerman (1990);
Greenfeld (1989); Howard (1985); Loewenberg (1984); Shaw (1990); Wuthnow and Shram
(1983).
13 For a discussion of social movement ideologies, see, for example, Laraña et al. (1994);
Oberschall (1993); Ryan (1992); Sassoon (1984).
14 Managerial ideologies are studied in Barley and Kunda (1992); Enteman (1993); Gren er
and Hogler (1991); Le Goff (1992); Miyajima (1986); Weiss (1986).
15 Feminist ideologies are studied in, for example, Ballaster (1991); Billington (1982);
Poole and Zeigler (1981); Ryan (1992); Sharistanian (1986).
16 Not only commonsense conceptions of racism, but also much scholarly work on racism,
rather exclusively identifies racism or 'racists (only) with right-wing, extremist groups, parties
or organizations. Despite the fact that racist is usually defined as an inter-group phenomenon
(e.g. between white Europeans and Others), white people us' are never racist. See, for
example, Able (1995); Blackwell (1994); Landau (1993); Thompson (1994). Indeed, as we
have seen before, also most leaders of racist parties will deny that they are racist, but at most
nationalist'.
17 For the strategies of the denial of racism, see van Dijk (1992).
18 For prototypes, see Rosch and Lloyd (1978).
Chapter 16
1 Many approaches to ideology implicitly or explicitly discuss social (group) relations as
the basis for ideology. Thus, in psychology, Social Dominance Theory (SDT) by Jim Sidanius
and his associates assumes that'humans are predisposed to form group-based social hier-
Notes
333
archies'. Moreover, especially people of greater status seem to display a greater tendency to
ingroup favouritism (Sidanius, 1993; Sidanius and Ekehammar, 1980, 1983; Sidanius et al.,
1994). Although 1 reject the (innate, natural?) disposition argument about social hierarchies,
my work on elite discourse and racism suggests indeed the special role of elites in the
reproduction of one type of ingroup favouritism: racism (van Dijk, 1993a; see also Sidanius
and Liu, 1992).
2 For details on power, see, for example, Clegg (1989); Lukes (1974, 1986); Oleson and
Marger (1993); Wrong (1979).
3 Persuasive power and mind control has been studied in several disciplines and from many
perspectives. From a socio-political perspective, namely, that of hegemony, the classical
source remains Gramsci (1971). A contemporary, more political orientation to the'manufacturing of consent' has been presented by Herman and Chornsky (1988). Mind control by the
media has been studied in a rich and controversial tradition of media power, 'influence' and
the 'effects' of mass communication (among numerous studies, see Altheide (1985); Altschull
(1984); Bryant and Zillmann (1986); Curran et al. (1987); Klapper (1960); Schiller (1973).
The cognitive and social psychological dimensions of mind control, usually defined as
'persuasion' or more critically as 'manipulation', have been studied by, for example, Bostrom
(1983); Bradac (1989); Harris (1989); Margolis and Mauser (1989).
4 See for instance Foucault's work on power (e.g. Foucault, 1980).
5 For discussion about the historical roots of racist ideologies and their relations to the slave
system, see, for example, Barker (1978).
Chapter 17
1 About the role of 'ideologues' in the formation of ideologies, see the contributions in Lau
and Sears (1986). My use of'entes' is complementary to the customary use in the social
sciences (e.g. Domhoff, 1978; Domhoff and Ballard, 1968; Milis, 1956). It especially
emphasizes, within the framework of critical discourse analysis, the special access to, and
control over public discourse, by the elites (see, e.g., van Dijk, 1993a, 1995). Lau et al. (1991)
provide some empirical evidence about the persuasiveness of policy proposals and of the role
of the elites in decision making. Zaller (1990) shows how ordinary citizens use elite cues in
order to transform their value orientations into support for speciflc policies. Jennings (1992),
finally, showed that party elites generally have more stable and consistent ideologies than
mass publics'.
2 For differences of access to public discourse, see van Dijk (1996). Generally it is found
that those who have more knowledge or expertise about politics, also have more consistent
attitudes and ideologies (see, e.g., Judd and Downing, 1990).
3 On ideologues, see, for example, Langston (1992); Martin (1983); Welch (1984). The role
of group leaders and the development of ideology has been studied by, e.g., Blommaert,1991;
Dreier, 1982; Folkertsma, 1988; Gaffney, 1989; Garcia, 1989).
4 For the everyday implications of feminist ideology, see, for example, Flaherty (1982);
Krishnan (1991); Redchift and Sinclair (1991); Ryan (1992); Sharistanian (1986); Togeby
(1995).
5 For feminist ideologies, see Note 4. For ethnically or 'racially' based ideologies of
resistance, see, for example, Fatton (1986); Innis and Feagin (1989); Marable and Mullings
(1994); McCarthney (1992); Turner and Wilson (1976). For environmental ideologies, see
Buttel and Flinn (1978).
6 For theoretical and empirical studies of elite racism, see van Dijk (1993a); for the way the
elites may frame racial issues, see also Kinder and Sanders (1990).
7 There is some experimental evidence on group-based dominance (e.g. of gender or' me')
that seems to support this assumption: For members of high status groups (e.g. whites, males)
there is a positive correlation between desire for group dominance and group affiliation (see,
e.g., Sidanius et al., 1994).
334
Notes
8 For details on the relations between ideologies, immigration policies and political
rhetoric, see, for example, Fitzgerald (1996).
Chapter 18
1 This discussion about dominant ideologies has been sparked especially by Abercrombie
et al. (1980, 1990). See also Howe (1994) for a critique of the critique of dominant ideologies
and for alternative suggestions to a Marxist concept of ideology, that is, as a 'possibly
contradictory set of themes'.
2 About these relations between the media and (other) power elites and their ideologies, see
Connell (1978); Dreier (1982); Fletcher (1991); Golding and Murdock (1979); Lichter et al.
(1990); Negrine (1989); Paletz and Entinan (1981); Rothman and Lichter (1984); Dreier
(1982).
3 See Hall (1980, 1982) for such an approach to the overall, ideological influence of the
media, which is also defended and illustrated in, for example, Herman and Chomsky (1988);
Schiller (1973); van Dijk (1991).
4 This again brings in the discussion of media effects and influence. That media in many
domains have power, is beyond dispute, and documented in many studies (see references in
Note 2). Whether and how exactly they have a pervasive, and not just a marginal or occasional
influence on the basic attitudes and ideologies of the reader, is much harder to prove (or to
reject). See for debate, for example, Bryant and Zillman (1986); Graber (1988); Iyengar and
Kinder (1987); Liebes and Katz (1993); MacKuen and Coombs (1981); Morley (1986,1993);
Neuman et al. (1992).
5 See Mannheim (1936). Note, though, that for Mannheim, utopias (like the ideologies of
dominant groups) are essentially misguided, because they are so 'strongly interested in the
descruction and transformation of a given condition of society' that their 'thinking is incapable
of correctly diagnosing an existing condition of society' (p. 36). Because, as suggested before,
both dominant as well as non-dominant ideologies may be true or false, also ideologies of
resistance are not necessarily based on valid analyses of the social order. Nor, however, is the
opposite necessarily false. Indeed, as suggested before, whereas dominant groups may have
interest in ignoring or denying true relations of domination in order to legitimate or conceal
their power, opposition groups should rather have a correct view of the social situation, in
order to be better placed to transform it. Thus, it would be quite inappropriate to generally
qualify feminist or anti-racist 'utopias' as misguided diagnoses of gender and yace domination
in society. In other words, the way dominant and oppositional groups understand the social
world cannot be the same (or similarly misguided). On the contrary, the ideological basis of
their beliefs is itself (also, though not only) a function of their respective social positions and
interests, and hence by definition different.
6 On the ideologies and popular success of Thatcherism and Reagonomics, see, for
example, Hall (1988); Kiewe and Houck (1991); Krieger (1986); Langston (1992). Yantek
(1988). For the New Right, see, for example, Bennett (1990); Levitas (1986); Sunic (1990).
7 On various forma of resistance and dissidence, see, for example, Fisher and Davis (1993);
Hall and Jefferson (1976); Luke (1989); Miller et al. (1989); Mullard (1985); Scott (1986);
Sivanandan (1982).
Chapter 19
1 Ideological institutions, also defined as 'ideological state apparatuses', and their role in
reproduction, have, for example, been discussed in more philosophical terms by Althusser
(1984); as well as in several books by Foucault (see, e.g., Foucault, 1972, 1979). From a
different perspectiva, much empirical work has been done on the ideological role of
organizations (Alvesson, 1987, 1991; Berezin, 1991; Downey, 1986; Goll and Zeitz, 1991;
Hill and Leighley,1993; Jones, 1984; Mumby, 1988; Sassoon, 1984; Theus, 1991; Weiss,
Notes
335
1986). Douglas (1986) specifically also focuses on organizations or institutions as 'thinking'
and hence as instances that develop ideologies, just as they make decisions.
2 The ideological functions of the family have been investigated more generally in much
primary socialization research, but also more specifically, for instance with respect to the
acquisition of gender roles, the acquisition of prejudices, and so on. See, for example, Aboud
(1988); Gittins (1993); Kraut and Lewis (1975); Todd (1985); Walsh (1983).
3 The ideological functions of schooling and formal education have been among the beststudied institutional aspects of ideology. See, among many other publications, the following
studies: Apple (1979, 1982); Apple and Weiss (1983); Ekehammar et al. (1987); Giroux
(1981); Karabel and Halsey (1977); Rothstein (1991); Sarup (1991); Sharp (1980); Stevens
and Wood (1992); Tiemey (1991); Watt (1994); Willis (1977); Young (1971). More
specifically, Baer and Lambert (1990) found that students of business and the professions tend
to support dominant ideologies, and those who studied social sciences tend consequently to
support counter-ideologies.
4 The ideological influence of the mass media has been discussed within the broader
framework of the power, effects and influence of the media, and has alternatively been
emphasized or mitigated, depending on theory and empirical findings. See Barrett et al. (1979);
Cormell (1978); Downing (1984); Fletcher (1991); Fowler (1991); Golding and Murdock
(1979); Hachten (1981); Hall (1982); Hartley and Montgomery (1985); Rothman and Lichter
(1985); Schiller (1973); Schiller and Alexandre (1992); Thompson (1990). For the relations
between news values and ideologies, see Westerstahl and Johansson (1994).
5 For the role of political panties and organizations in the reproduction of racist ideologies,
see, for example, Ben-Tovim et al. (1986); Browning et al. (1990); Feldman (1992); Fitzgerald
(1996); Kinder and Sears (1981); Lauren (1988); Layton-Henry (1992); Miles and Phizacklea
(1979); Reeves (1983); Sniderman et al. (1993); Solomos (1986, 1993); van Dijk (1993a).
6 For the role of science and scholarship in the reproduction of racism, see, for example,
Barkan (1992); Benedict (1982); Chase (1975); Essed (1987); Haghighat (1988); Joseph et al.
(1990); Shipman (1994); Tucker (1994); Unesco (1993); van Dijk (1993a).
Chapter 20
1 This does not mean, of course, that diere is no earlier work on discourse and ideology.
See the references given in Note 4. The problem is that much work on discourse and ideology
does not discuss discourse structures in any detall at all, or vaguely identifies discourse with
ideology. Though critical of such identification, Purvis and Hunt (1993) simply continue the
reduction of discourse to some kind of overall'order of discourse' without actually analysing
it, thus continuing the (usually Marxist or Foucauldian) tradition they criticize. See, for
example, Fairclough (1992) for a discourse analytical critique of Foucault.
2 This is the approach especially advocated, in various degrees of orthodoxy, by discursive
psychologists in the UK, an approach we have commented on before. See, for example, Potter
and Wetherell (1987, 1989); Billig (1991b). For discussion of this approach, see also
Augoustinos and Walker (1995).
3 Although crucially relevant for many forms of communication and interaction, we must
unfortunately ignore this broader 'seiniotic' approach to discourse in this book. For ideological
implications of various types of visual communication, see, for example, Austin (1977);
Barker (1989); Hall et al. (1980); Hodge and Kress (1988); Kress and Van Leeuwen (1990);
Pauly (1993); Reis (1993); Shohat and Stam (1994).
4 Implicitly, many studies of the acquisition or expression of ideology deal with language,
discourse or communication. Focused interest in the role of discourse in the acquisition and
change of ideologies can be found in the following studies, among many others: Aronowitz
(1988); Barley and Kunda (1992); Billig (1991b); Boylan and Foley (1992); Burton and Carlen
(1979); Dant (1991); Fairclough (1989, 1995); Fowler (1991); Hodge and Kress (1993);
Mumby (1988); Pecheux (1982); Reis (1993); Rossi-Landi (1978); Strassner (1987); van Dijk
(1995); Wenden and Schaffner (1994); Wodak (1989, 1996); Wuthnow (1989). However, very
336
Notes
few of these studies provide a detailed and systematic study of the relations between
ideological structures and discourse structures.
5 See, for example, Foucault (1981).
6 For this critical approach, critical discourse analysis (CDA), see for instance, van Dijk
(1993b). Of course, diereis much other work in discourse analysis on ideology (see Note 3),
and other directions of CDA. However most of diese only bridge the gap between (linguistic
and other) approaches to discourse structures on the one hand, and social interaction or social
structure, on the other, and neglect the important cognitive 'interface'. Similarly, important
work on the cognitive psychology of discourse production and comprehension, usually
neglects the social basis of discourse and understanding. One of the few approaches in critical
discourse analysis that integrates these different dimensions is the work of Ruth Wodak and
her associates (see, e.g., Wodak, 1987, 1989, 1991, 1996). Finally, much of the tradition of
critical linguistics in the UK and Australia (such as the work, cited in Note 3, by Fowler,
Kress, Van Leeuwen and others), has been formulated in the broader framework of functional,
systemic linguistics and semiotics as initiated by Halliday (1973, 1985, 1987).
7 For other approaches to discourse, see the following introductions: Renkema (1993);
Schiffrin (1993); van Dijk (1985, 1997).
8 We shall not engage in this debate here, nor detall the many differences between our
framework and the more philosophical or postmodem approaches to discourse. For discussion,
see, for example, Agger (1990, 1992,1993); Rojek and Turner (1993); Simons and Billig
(1994); see also Fairclough' s assessment of the relevance of Foucault for discourse analysis
(Fairclough, 1992).
9 See the references given in Note 2.
10 A typical example of more contemporary 'social' semiotic analysis that integrates
modem linguistics and discourse analysis, is the work by Hodge and Kress (1988) and Kress
and Van Leeuwen (1990).
Chapter 21
1 For analyses of graphical or visual properties of discourse, see, for example, Hodge and
Kress (1988); Kress and Van Leeuwen (1990); Mitchell (1994); Rutter (1984); Solso (1994);
Saint-Martin (1990).
2 The ideological implications of visual communication have been studied by, for
example, Austin (1977); Bristor et al. (1995); Davis and Walton (1983); Doise (1978);
Ellsworth and Whatley (1990); ElWarfally (1988); Mitchell (1986); Pauly (1989); Sinclair
(1987).
3 For the study of sound structures in discourse, and especially of intonation, see Brazil
(1983); Gibbon and Richter (1984); Selting (1995).
4 Of the few studies that relate phonological variables with social or political functions,
see Moosmüller (1989); Van Leeuwen (1992). In conversation analysis, special attention is
paid to the nature and functions of applause. A politically oriented study of applause is given
by Atkinson (1984).
5 An early study of die role of syntax in the expression of ideologically based meanings is
Fowler et al. (1979).
6 For the ideologically based expression of agency and responsibility, see, for example,
Fowler (1991); Fowler et al. (1979); Sykes (1985); van Dijk (1991); van Leeuwen (1995).
7 For social, political and ideological studies of pronouns, see for instance Brown and
Gilman (1960); Carbó (1987); Duranti (1984); Jacquemet (1994); Maitland and Wilson
(1987); Urban (1988); van Dijk (1987); Wilson (1990). For experimental evidence about the
persuasiva role of the use of 'Us' and 'Thern' in social categorization, sea Perdue et al.
(1990).
8 For ideological studies of discourse meanings, see, for example, Luke (1989); Pecheux
(1982); van Dijk (1995).
9 For news schemata and their ideological implications, see van Dijk (1988a, 1998b).
Notes
337
10 The ideological implications of the use of rhetorical figures (and especially metaphor)
have been studied in much work, such as in Billig (1991b, 1995); Billig and Sabucedo (1994);
Chilton (1995); Gale (1994); Kenshur (1993); Lakoff (1987, 1995); Medhurst (1990); Miller
and Fredericks (1990); Montgomery et al. (1989); Mumby and Spitzack (1983); Roeh and Nir
(1990); van Dijk (1991); Wander (1984).
11 Ideological analysis of conversational structures and strategies is a direction of research
that until recently was anathema in most conversation analysis. However, there are now
several studies of conversation that focus on social relations that may have an ideological
basis, such as those of gender or profession. See, for example, Atkinson (1984); Boden and
Zimmerman (1991); Firth (1995); Greatbatch (1992); Heritage (1985); Heritage and Greatbatch (1986); West (1979, 1984, 1990).
Chapter 22
1 Despite the absence of a general theory of context, there have been many writers,
especially also in the ethnography of discourse, who have dealt with various aspects of
context. See, for example, the well-known SPEAKING